冷冻和解冻速率对水溶液中蛋白质变性的影响

yanyi8888 · 发表于 2021-12-14 15:34:17

欢迎您注册蒲公英

您需要登录才可以下载或查看，没有帐号？立即注册

x

Effect of Freezing and Thawing Rates on Denaturation of Proteins in

Aqueous Solutions

冷冻和解冻速率对水溶液中蛋白质变性的影响

Enhong Cao,1 Yahuei Chen,1Zhanfeng Cui,1 Peter R. Foster2

1Department of EngineeringScience, University of Oxford, Parks Road, Oxford OX1 3PJ, United Kingdom

2Protein Fractionation Centre, ScottishNational Blood Transfusion Service, Edinburgh, United Kingdom

Received 11 July 2002; accepted 7 November 2002 DOI:10.1002/bit.10612

Abstract: The freeze denaturation of model proteins, LDH, ADH, andcatalase, was investigated in absence of cryoprotectantsusing a microcryostage under well- controlled freezing and thawing rates.Most of the experi- mental data wereobtained from a study using a dilute solutionwith an enzyme concentration of 0.025 g/l. Thedependence of activity recovery of proteins on the freez- ing and thawing rates showed areciprocal and indepen- dent effect, that is, slow freezing (at a freezing rateabout 1°C/min) and fast thawing (at athawing rate >10°C/min) produced higher activity recovery, whereas fastfreezing with slow thawing resulted in more severe damage to proteins. With minimizing the freezingconcentration and pH change ofbuffer solution by using a potassium phos- phate buffer, this phenomenon couldbe ascribed to sur- face-induceddenaturation during freezing and thawing process.Upon the fast freezing (e.g., when the freezing rate >20°C/min), small icecrystals and a relatively large surfacearea of ice–liquid interface are formed, which in- creases the exposure of protein molecules to the ice– liquidinterface and hence increases the damage to theproteins. During thawing, additional damage to proteins is caused byrecrystallization process. Recrystallization exerts additional interfacialtension or shear on the en- trappedproteins and hence causes additional damage tothe latter. When buffer solutes participated during freez- ing, the activity recovery ofproteins after freezing and thawing decreased due to the change of buffersolution pH during freezing. However, the patterns of the depen- dence onfreezing and thawing rates of activity recovery did not change except for that atextreme low freezing rates (<0.5°C/min). The results exhibited that the freezing damage of protein in aqueoussolutions could be re- duced by changing the buffer type and composition and byoptimizing the freezing–thawing protocol. © 2003 Wiley Periodicals, Inc. Biotechnol Bioeng 82: 684–690, 2003.

Keywords: protein freezing; freeze denaturation;freezing

rate; thawing rate; buffer type; lactatedehydrogenase; cryomicroscope

摘要：模型蛋白、LDH、ADH 和过氧化氢酶的冷冻变性在没有冷冻保护剂的情况下使用微冷冻台在良好控制的冷冻和解冻速率下进行了研究。大多数实验数据是从使用酶浓度为 0.025 g/l 的稀释溶液的研究中获得的。蛋白质的活性恢复对冻融速率的依赖性表现出相互的和独立的效应，即缓慢冻结（以约 1°C/min 的冻结速率）和快速解冻（以解冻速率 > 10°C/min) 产生更高的活性恢复，而快速冷冻缓慢解冻导致更严重的蛋白质损伤。通过使用磷酸钾缓冲液来最小化缓冲溶液的冷冻浓度和 pH 值变化，这种现象可以归因于冷冻和解冻过程中的表面诱导变性。在快速冷冻时（例如，当冷冻速度 >20°C/min 时），会形成小冰晶和相对较大的冰液界面表面积，这增加了蛋白质分子对冰液的暴露界面，从而增加对蛋白质的损伤。在解冻过程中，再结晶过程会对蛋白质造成额外的损害。重结晶会对被捕获的蛋白质施加额外的界面张力或剪切力，从而对后者造成额外的损害。当缓冲溶质在冷冻过程中参与时，由于冷冻过程中缓冲溶液pH值的变化，冷冻和解冻后蛋白质的活性恢复降低。然而，除了极低的冷冻速率（<0.5°C / min）外，对活性恢复的冷冻和解冻速率的依赖模式没有改变。结果表明，可以通过改变缓冲液类型和组成以及优化冻融方案来减少水溶液中蛋白质的冷冻损伤。 © 2003 Wiley Periodicals, Inc.Biotechnol Bioeng 82: 684–690, 2003。

关键词：蛋白质冷冻；冷冻变性；冷冻速度;解冻率；缓冲区类型；乳酸脱氢酶;冷冻显微镜

Correspondenceto: Professor Zhanfeng Cui, Department of EngineeringScience, University of Oxford, Parks Road, Oxford OX1 3PJ, United Kingdom.E-mail: [url=]zhanfeng.cui@eng.ox.ac.uk[/url]

通讯作者：英国牛津大学工程科学系，Parks Road, Oxford OX1 3PJ，ZhanfengCui 教授。电子邮箱：zhanfeng.cui@eng.ox.ac.uk

Contractgrant sponsor: Biotechnology and Biological Science Research Council (BBSRC) ofthe U.K.

合同资助方：英国生物技术和生物科学研究委员会 (BBSRC)

INTRODUCTION介绍

Freezing or freezing–drying is common unit operation in the production of therapeutic proteins andenzymes as well as cell biomass.However, freezing can induce several stressesthat are capable of denaturing proteins (Franks, 1985), such as cold temperature, ice formation,solute concentration due to thecrystallization of water, eutectic crystallization of buffer solutes, and resultant pH changes. Much research has beenreported (Anchordoquy and Carpenter, 1996; Arkawaet al., 1993; Tamiya et al., 1985) emphasizing the employ- ment ofdifferent cryoprotectants to increase the thermody- namic stability of thenative state of the protein. The use of theprotectants can sometimes be problematic in affecting the quality and purity of the products. 冷冻或冷冻干燥是治疗性蛋白质和酶以及细胞生物质生产中的常见单元操作。然而，冷冻会引起几种能够使蛋白质变性的压力 (Franks, 1985)，例如低温、结冰、水结晶导致的溶质浓度、缓冲溶质的共晶结晶以及由此产生的 pH 值变化。许多研究报告（Anchordoquy 和 Carpenter，1996年；Arkawa 等人，1993 年；Tamiya 等人，1985 年）强调使用不同的冷冻保护剂来增加蛋白质天然状态的热力学稳定性。保护剂的使用有时会影响产品的质量和纯度。

Quick freezing of enzymes in liquid nitrogen was often recommended in order to preserve theenzymatic activity (Nema, 1993;Shikama and Yamazaki, 1961), but several studieshave reported that rapid freezing in liquid nitrogenfailed to prevent some proteins from freezing denaturation (e.g., Eckhardt et al., 1991; Schwartz and Batt,1980). More recent studies onfreezing of protein solutions (Chang et al., 1996; Jiang and Nail, 1998;Strambini and Gabellieri, 1996) showed that fast freezingproduced more damageto proteins and gave lowerrecovery of activity after freezing and thaw-ing, the phenomenon being explained as the ice-induced partial unfoldingof proteins. However, there is a lack of systematicstudy of the freezing rate dependence of protein denaturation and very littlework (e.g., Jiangand Nail, 1998) hasaddressed how to reduce freezing damage during freez- ing and thawing throughthe optimization of operation con- ditions.

通常建议在液氮中快速冷冻酶以保持酶活性（Nema，1993；Shikama 和 Yamazaki，1961），但一些研究报告称，在液氮中快速冷冻未能防止某些蛋白质冷冻变性（例如 , Eckhardt 等人，1991 年；Schwartz 和 Batt，1980 年）。最近关于蛋白质溶液冷冻的研究（Chang 等人，1996 年；Jiang 和Nail，1998 年；Strambini 和Gabellieri，1996 年）表明，快速冷冻对蛋白质产生更多的损害，并且在冷冻和解冻后活性恢复率较低，这种现象被解释为冰诱导的蛋白质部分解折叠。然而，缺乏对蛋白质变性的冷冻速率依赖性的系统研究，并且很少有工作（例如，Jiang and Nail，1998）解决了如何通过优化操作条件来减少冻融过程中的冻害。版本。

In the present study, the effect of the freezing andthaw- ing rates on the activity of enzymes in aqueous solutions wassystematically examined in the absence of cryopro- tectants. A cryomicrostagesystem was employed to provide aprecise control of freezing and thawing rates. Three en- zymes, LDH, ADH, and catalase, were used as model pro- teins.In addition to the cooling and thawing rates, the effect of buffer conditions was also experimentally studied.This study aimed at obtaining a better understanding on the dif- ferentmechanisms for protein denaturation in the freezing and thawing process andproviding a guidance in the selec- tion and optimization of freezing andthawing operation conditions.

在本研究中，在没有冷冻保护剂的情况下系统地检查了冷冻和解冻速率对水溶液中酶活性的影响。采用低温微量级系统来精确控制冷冻和解冻速率。三种酶，LDH、ADH 和过氧化氢酶，用作模型蛋白。除了冷却和解冻速率外，还通过实验研究了缓冲条件的影响。本研究旨在更好地了解冻融过程中蛋白质变性的不同机制，为冻融操作条件的选择和优化提供指导。

MATERIALS AND METHODS 材料和方法

Materials and Preparation 材料和准备

Lactatedehydrogenase (LDH) (Type II, from rabbit muscle), alcohol dehydrogenase (ADH)from baker’s yeast, catalase from bovine liver, and the bovine serum albumin(BSA) were purchased from Sigma Chemical (St. Louis, MO). The LDH product, acrystalline suspension in 3.2 M (NH4)2SO2 solution, was used directly in the experimentwithout removing the ammonia sulfate. Five buffer solu- tions were employed inthis study: potassium phosphate, sodium phosphate, citrate, HEPES, andTris-HCl. All chemicals were reagent grade and obtained from Sigma Chemical.

乳酸脱氢酶 (LDH)（II型，来自兔肌肉）、来自面包酵母的乙醇脱氢酶 (ADH)、来自牛肝的过氧化氢酶和牛血清白蛋白 (BSA) 购自 Sigma Chemical（密苏里州圣路易斯）。 LDH 产物是 3.2 M (NH4)2SO2 溶液中的结晶悬浮液，直接用于实验中，无需去除硫酸铵。本研究中使用了五种缓冲溶液：磷酸钾、磷酸钠、柠檬酸盐、HEPES 和Tris-HCl。所有化学品均为试剂级，购自Sigma Chemical。

The stock protein solution was prepared by dissolving a certainamount of well-mixed enzyme suspension or powder in 1 ml of buffer solution to forma protein concentration of approximately1 mg/ml, which was incubated for 40 min at 37°C to ensure dissolution of crystals and powders. Sequen- tial assays, carried out at10-min intervals, showed that the enzymesolutions reached full activity after 30 min and did not change for 2 days. The protein concentration of the stock solution was determined by usingthe Bradford re- agent, with BSAused as standard. The stock solution was furtherdiluted in one or two steps to the final concentrationdesired. The diluted solutions exhibited about a 10% de- crease in activity but were stablethereafter for a few hours.

储备蛋白溶液是将一定量混合好的酶悬液或酶粉溶解在1ml缓冲液中，使蛋白浓度约为1mg/ml，37℃温育40min保证溶解。晶体和粉末。以 10 分钟间隔进行的连续测定表明，酶溶液在 30 分钟后达到完全活性，并且在 2 天内没有变化。储备溶液的蛋白质浓度使用Bradford 试剂测定，BSA 用作标准。将储备溶液分一或两步进一步稀释至所需的最终浓度。稀释后的溶液活性降低了约 10%，但此后数小时内保持稳定。

Enzyme Activity Assays 酶活性测定

LDHwas assayed at 25°C using the procedure described by Worthington Biochemical Corporation (Lakewood, NJ) us- ing a spectrophotometer (UNICAM uv4series). The assay is based on theconversion of pyruvate to lactate. During the reaction, an equimolar amount of NADHand pyruvate is oxidized to NAD. Theoxidation of NADH results in a decreasein absorbance at 340 nm. The reaction mixture consisted of 2.8 ml of 0.2 M Tris-HCl buffer (pH 7.3), 0.1 ml of 30 mM sodium pyruvate, and 0.1 ml of 6.6 mM NADH. The reactionwas initiated by the addition of 0.1 ml ofproperly diluted enzyme solution. The enzyme activity was evaluated based onthe measured decrease in absor- bance at 340nm.

LDH 在 25°C 下使用 Worthington Biochemical Corporation(Lakewood, NJ) 描述的程序使用分光光度计（UNICAMuv4 系列）进行测定。该测定基于丙酮酸向乳酸的转化。在反应过程中，等摩尔量的 NADH 和丙酮酸被氧化成 NAD。 NADH 的氧化导致 340 nm 处的吸光度降低。反应混合物由 2.8 ml 0.2 M Tris-HCl 缓冲液 (pH 7.3)、0.1ml 30 mM 丙酮酸钠和 0.1 ml 6.6 mM NADH 组成。通过加入0.1ml适当稀释的酶溶液引发反应。根据在 340 nm 处测得的吸光度降低来评估酶活性。

ADH activity was measured using the protocol provided bySigma, based on the reduction of NAD by ethanol to NADH, which results in anincrease in absorbance at 340 nm.The reaction mixture consisted of 0.05 M sodiumpy- rophosphate (pH 8.8) 1.3 ml, 95%(v/v) ethanol 0.1 ml, and 0.015 M -NAD 1.5 ml. The reaction was started by theaddition of 0.1 ml of properly diluted enzyme solution. The enzyme activity wascalculated from the measured increase in absorbance at 340 nm.

ADH 活性是使用 Sigma 提供的方案测量的，基于乙醇将 NAD 还原为 NADH，这导致 340 nm 处的吸光度增加。反应混合物由 0.05 M 焦磷酸钠 (pH 8.8) 1.3 ml、95%(v/v) 乙醇 0.1 ml 和 0.015 M -NAD 1.5ml 组成。通过加入0.1ml适当稀释的酶溶液开始反应。根据在 340 nm 处测得的吸光度增加来计算酶活性。

Catalase activity assay was conducted using the proce-dure provided by Sigma, based on the first-order decompo- sition of peroxide bycatalase. The reaction was initiated by the addition of 0.1 ml of dilutedcatalase solution to 2.9 ml of 0.036% (w/w) hydrogen peroxide in 0.05 M potassium phosphate buffer (pH 7.0).The catalase activity was deter- mined from the measurement of absorbance at240 nm.

基于过氧化氢酶对过氧化物的一级分解，使用Sigma 提供的程序进行过氧化氢酶活性测定。通过将 0.1 ml 稀释的过氧化氢酶溶液添加到 2.9 ml 0.05 M 磷酸钾缓冲液 (pH 7.0) 中的 0.036% (w/w) 过氧化氢中来引发反应。过氧化氢酶活性通过测量 240 nm 处的吸光度来确定。

Cryomicroscope System 冷冻显微镜系统

Thecryomicroscope system consists of an Olympus BX-50microscope and a Linkam BCS 196 cryostage (Linkam Sci- entific, Tadworth, Surrey, U.K.). Thecryostage consists of a silverblock, 22 mm in diameter, and a temperature moni-tor and control system. Liquid nitrogen (LN2) is introduced directlyfrom a 2-liter Dewar through a small-bore tube, via a filter, by a cooling pump (Linkam LNP). The heating elementis cast into the silver block to form an integral heater and sensor assembly. Aplatinum resistor mounted near the surface of the cryostage provides a stableand ac- curate temperature signal.A small quantityof vaporized gas is recycled from the pump and passedover the top surface of the cryostage window to prevent condensation. Samplecarriers have been designed to hold either a 16-mm cover slip for freezing a thin-layer sample, or a 16-mm quartzcrucible for freezing large samples, which can be loaded on to the silver block from the top of thestage by removing the lid or from the side of the stage via the side door.

冷冻显微镜系统由 Olympus BX-50 显微镜和 Linkam BCS 196 冷冻台（Linkam Science, Tadworth, Surrey, U.K.）组成。低温台由一个直径为 22 mm 的银块和一个温度监测器和控制系统组成。液氮 (LN2) 直接从 2 升杜瓦瓶中通过小口径管、过滤器、冷却泵 (Linkam LNP) 引入。加热元件被浇铸在银块中，形成一个整体的加热器和传感器组件。安装在冷冻台表面附近的铂电阻器可提供稳定和准确的温度信号。少量汽化气体从泵中回收并通过低温级窗口的顶部表面以防止冷凝。样品架设计用于固定用于冷冻薄层样品的 16 毫米盖玻片，或用于冷冻大样品的 16 毫米石英坩埚，可以通过以下方式将其从载物台顶部装载到银块上取下盖子或通过侧门从舞台侧面取下。

The BCS 196 cryostage is directly mounted on the mi- croscope substage for stability andallows standard objec- tive lensesto be used. All the essential controls for the operation of the temperatureprogrammer are performed by Link 2.0Programmer Control software. A real time videoand measurement system (RTVMS) was used to display and torecord the real-time video through a video camera (JVC TK C1381).

BCS 196 低温载物台直接安装在显微镜载物台上以确保稳定性，并允许使用标准物镜。温度编程器操作的所有基本控制均由 Link 2.0 Programmer Control 软件执行。实时视频和测量系统 (RTVMS) 用于通过摄像机 (JVC TK C1381) 显示和记录实时视频。

Before the experiment, the temperature accuracy of the cryostage and the temperaturedifference between the stage and thelarge volume of sample (55 l) were examined. Approximately 3 l of aqueoussolution of NaCl (23.3% wt) or KCl (19.75% wt) was pipetted to thesample crucible at room temperature, and a 13-mm glass cover slide was placed over the sample to form asandwich. The sample was first cooled at a rate of 40°C/min to−45°C, held for 2 min, and was then warmed at a rate of 40°C/min to about 5°C below the eutectic temperature of thesample. The further heating in 1°C/min was curried out manually until eutecticmelting was observed. The temperature indicated was re- corded compared with the eutectic point of the solution. It was found that at the eutectic meltingpoints, the indicated (cryostage) temperatures were very close to the values of eutectic melting of the samples.

实验前，对冷冻台的温度精度以及冷冻台与大体积样品（55 l）之间的温差进行了检查。在室温下将大约 3 l NaCl (23.3% wt) 或 KCl (19.75% wt) 的水溶液移液到样品坩埚中，并将 13 毫米玻璃盖玻片放在样品上以形成三明治。首先将样品以 40°C/min 的速率冷却至 -45°C，保持 2 分钟，然后以 40°C/min 的速率升温至样品共晶温度以下约 5°C . 以 1°C/min 的速度进一步加热手动进行，直到观察到共晶熔化。记录的温度与溶液的共晶点进行比较。发现在共晶熔点下，指示的（低温阶段）温度非常接近样品的共晶熔化值。

For measuring the temperature difference between thestage and sample, 55 l of buffer solution was loaded on to the quartz crucible.A calibrated fine wire thermocouple (T-type)was embedded in the buffer solution. The samplewas cooled in rates of 1–60°C/min from room temperature to −20 to −40°C,respectively. The data from the thermo- couple were collected by a Datalogger(HP34970A, Ampli- con Liveline Ltd., Brighton, U.K.) and used to comparewith these from the cryostage readout. The difference between the stage target temperature and the sample temperature was found to be quite close, and moreimportantly the sample cooling rate is identical to that of the indicated.

为了测量载物台和样品之间的温差，将 55 升缓冲溶液加载到石英坩埚上。校准好的细线热电偶（T 型）嵌入缓冲溶液中。将样品以 1–60°C/min 的速率分别从室温冷却至 -20 至-40°C。来自热电偶的数据由数据记录器（HP34970A，AmpliconLiveline Ltd.，Brighton，英国）收集，并用于与来自低温级读数的数据进行比较。发现载物台目标温度和样品温度之间的差异非常接近，更重要的是样品冷却速率与指示的相同。

Freezing–Thawing冻融

Thecooling–warming protocol was made up by defining 3 ramps: (1) Pro-cooling: the samples were pre-cooled at30°C/min to 0°C and held for 2 min to obtain same initial temperature conditions. (2) Freezing: the samples werecooled at a desired cooling (or freezing) rate to a target temperature and thenheld for 2 min to achieve uniform temperature.The rate of temperature change in this ramp isthe freezing rate as the actual solidification, with seeding, occurredat this temperature change rate, controlled by the cryostage controller. (3) Thawing: the samples were thawed at a designed warming (or thawing)rate to 10°C to ensure the sample iscompletely thawed. Similarly the rate of warming in this ramp is the thawingrate.

冷却-升温协议由定义 3 个斜坡组成：（1）预冷却：将样品以 30°C/min 的速度预冷至 0°C，并保持 2 分钟以获得相同的初始温度条件。 (2)冷冻：将样品以所需的冷却(或冷冻)速率冷却至目标温度，然后保持2分钟以达到均匀温度。该斜坡中的温度变化速率是冷冻速率，因为实际凝固和晶种发生在该温度变化速率下，由低温阶段控制器控制。 (3)解冻：以设计的升温（或解冻）速率将样品解冻至10°C，以确保样品完全解冻。类似地，该斜坡中的升温速率是解冻速率。

Unless otherwise stated, freezing and thawing experi-ments were performed using a quartz crucible. A protein concentration of 0.025 mg/ml was selected for mostexperi- mental work in this study following a preliminary investi- gation (seenext section); 55 l of protein solution wastransferred to the crucible which was then covered with a glass slip over the top for the bettertemperature accuracy. After a predefined freezing–thawing cycle, 50 l of thawedsample was collected and kept in an ice–water mixture and assayed within a maximum of 10 min. No significant changein enzyme activity was observed during this period. Non-frozen controls containingthe same amount of the pro- tein solution was kept in the sameice–water mixture and assayed before and after each experiment. Activity recovery was expressed as a percentage ofthe activity prior to freez- ing.

除非另有说明，冷冻和解冻实验均使用石英坩埚进行。在初步调查（见下一节）后，本研究中的大多数实验工作选择了 0.025 mg/ml 的蛋白质浓度；将55 l 蛋白质溶液转移到坩埚中，然后在坩埚顶部覆盖玻璃粉，以获得更好的温度精度。在预定义的冻融循环后，收集 50 升解冻样品并将其保存在冰水混合物中，并在最多 10 分钟内进行分析。在此期间未观察到酶活性的显着变化。将含有等量蛋白质溶液的非冷冻对照保存在相同的冰水混合物中，并在每次实验前后进行检测。活性恢复以冷冻前活性的百分比表示。

In the case of seeding, samples were pre-cooled to 1°C below the freezing point of thesolution and held for 2 min in thefirst ramp. A thin glass hollow sample tube (0.3 mm inside diameter) with a fine sealed tip was held over the surface of liquid nitrogen for ~30 s. The thin glass tube was quickly moved to the surface of the sample through the opened cryostage window at the end ofthe 2-min holding time and wasbrought in contact with the sample with the cooled tip. The secondramp was followedthereafter and the stagewindow was closed as quickly as possible.

在接种的情况下，样品被预冷至溶液冰点以下 1°C，并在第一个斜坡中保持 2 分钟。将带有精细密封尖端的细玻璃空心样品管（内径 0.3 毫米）保持在液氮表面上  30 秒。在 2 分钟保持时间结束时，将细玻璃管通过打开的冷冻台窗口快速移动到样品表面，并用冷却的尖端与样品接触。随后进行第二个斜坡，并尽快关闭舞台窗口。

RESULTS AND DISCUSSION 结果和讨论

Preliminary Observations 初步观察

Theeffects of protein concentration and freezing rate were firstly examinedwithout seeding. The dependence of freeze

denaturationon enzyme concentration in 50 mM potassiumphosphate buffer (pH 7) is shown in Fig. 1. The freezingand thawing protocols aregiven in the figure. For all three en- zymesstudied, significant activity loss (i.e., activity recov- ery < 100%) wasobserved at low concentrations, and ac- tivityrecovery was increased with the increase of enzyme concentrations. It was alsofound that the activity recovery is nearly 100% for enzyme solutions with aconcentrations of 1 mg/ml (results not shown in the figure). The increasedactivity recovery of proteins with increasing concentration during freezing andthawing has long been recognized (Fishbein and Winkert, 1977; Izutsu et al.,1994). The effect of concentration was believed to come from the poorer sta- bility of the proteins at the lowerconcentrations (Tamiya et al.,1985), but the mechanism has not been systematically investigated. It has been reported that the protective effect at lower protein concentrations can beachieved by the addi- tion ofmacromolecules such as BSA, PVP, polyethyleneglycol, etc. (Anchordoquy and Carpenter, 1996), which are preferentially excluded from thesurface of proteins and increase the overall thermodynamicstability of the native proteins.

首先在不接种的情况下检查蛋白质浓度和冷冻率的影响。冷冻变性对 50 mM 磷酸钾缓冲液 (pH 7) 中酶浓度的依赖性如图 1 所示。图中给出了冷冻和解冻方案。对于所研究的所有三种酶，在低浓度下观察到显着的活性损失（即活性恢复 < 100%），并且活性恢复随着酶浓度的增加而增加。还发现浓度为 1 mg/ml 的酶溶液的活性恢复率接近 100%（结果未在图中显示）。长期以来，人们已经认识到蛋白质在冷冻和解冻过程中随着浓度的增加而增加的活性恢复（Fishbein 和Winkert，1977 年；Izutsu 等人，1994 年）。浓度的影响被认为是由于蛋白质在较低浓度下的稳定性较差（Tamiya 等，1985），但尚未系统地研究该机制。据报道，通过添加 BSA、PVP、聚乙二醇等大分子可以实现较低蛋白质浓度下的保护作用（Anchordoquy 和 Carpenter，1996），这些大分子优先被排除在蛋白质表面之外。并增加天然蛋白质的整体热力学稳定性。

On the basis of the resultsshown in Fig. 1, a protein concentration of 0.025 mg/ml was chosen for the restof the experimental work. Proteinsat this concentration are sensi- tive to freezing rate, and this concentrationis also suffi- ciently high to minimize errors in solution transfer and han-dling, e.g., due to protein adsorption onto the surface of pipette tips.

在图 1 所示结果的基础上，选择 0.025 mg/ml 的蛋白质浓度用于其余的实验工作。该浓度的蛋白质对冷冻速率很敏感，而且该浓度也足够高，以最大限度地减少溶液转移和处理中的错误，例如，由于蛋白质吸附到移液器吸头表面。

Therecovery of activity of LDH in potassium phosphate buffer after freezing atdifferent rates was then examined under the conditions without seeding at fixedthawing rates, the results were illustrated in Fig. 2. A slight downward trendin activity recovery at higher applied freezing rates (when >10°C/min) wasobserved, but there was no statisti- cally significant dependence of freezedenaturation on the applied freezing rate. Lower recovery of protein activitywas observed when slow thawing (1°C/min) was applied.

然后在没有以固定解冻速率播种的条件下检查以不同速率冷冻后磷酸钾缓冲液中 LDH 活性的恢复，结果如图 2 所示。在较高应用冷冻速率下活性恢复略有下降趋势（当>10°C/min 时）观察到，但冷冻变性对应用的冷冻速率没有统计上显着的依赖性。当应用缓慢解冻 (1°C/分钟) 时，观察到蛋白质活性的恢复较低。

file:///C:/Users/ADMINI~1/AppData/Local/Temp/msohtmlclip1/01/clip_image002.gif

Figure 1. Effect ofconcentration on activity recovery of LDH, ADH, and catalase after afreezing–thawing cycle (protocol: applied freezing rate 10°C/min to −30°Cwithout seeding, thawing rate 15°C/min).

file:///C:/Users/ADMINI~1/AppData/Local/Temp/msohtmlclip1/01/clip_image004.gif

Figure2. Effect of the applied freezing rate on theactivity recovery of LDH (potassium phosphate buffer, without seeding).

It should be pointed out that the applied freezingrate (i.e., cooling rate in the second ramp) is not the actualfreezing rate when freezing without seeding due to subcooling. It was observed during the examination ofthe temperature measurement of the cryostage, that subcooling became pro- nounced as the cooling ratedecreased. Spontaneous nucle- ation occurredat −6 to −8°C when the coolingrate exceeded 30°C/min, butnot until −12 to −14°C when the cooling rate was 1°C/min. All samples exhibiteda freeze exotherm at thesupercooling point, and the sample temperature rose instantly to ~1°C then dropped to the temperature close to that of the stage with most of theliquid solidified, as ob- served onthe monitor. If dividing this temperature differ-ence by the time interval, the actual rates of freezing in samples in the initial solidificationprocess were found to be

−47.5,−61.4, −70.0, and −100.0°C/min corresponding tothe stage cooling rates of −1, −10, −30, and −60°C/min, respectively.The lower the applied cooling rate, the higher the subcooling, and the lower the actual nucleation tempera- ture. This larger solidification drivingforce results in a more rapid phase change when it occurs.The actual “freezing rate” in random nucleation freezing, i.e., withoutseeding, is much higher that theapplied cooling rate, and not so sen- sitiveto the latter. Hence the activity recovery is not so sensitive to the applied cooling rate, either.

应该指出的是，应用的冷冻速率（即第二个斜坡中的冷却速率）不是由于过冷而不播种而冷冻时的实际冷冻速率。在检查低温阶段的温度测量期间观察到，随着冷却速率的降低，过冷变得明显。当冷却速度超过 30°C/min 时，自发成核发生在 -6 到 -8°C，但当冷却速度为 1°C/min 时，自发成核发生在 -12 到-14°C。所有样品在过冷点都表现出冻结放热，样品温度立即上升到1°C，然后下降到接近于大部分液体凝固的阶段的温度，如在监视器上观察到的。如果将该温差除以时间间隔，则发现初始凝固过程中样品的实际冻结速率为-47.5、-61.4、-70.0和-100.0°C/min，对应于阶段冷却速率分别为 -1、-10、-30和 -60°C/min。应用的冷却速率越低，过冷度越高，实际成核温度越低。这种较大的凝固驱动力会导致发生更快速的相变。随机成核冷冻（即无晶种）中的实际“冷冻速率”远高于应用的冷却速率，并且对后者不那么敏感。因此，活性恢复对应用的冷却速率也不那么敏感。

It is hence revealed the importance of seeding. Seedingmust be applied in order to achieve a controllable freezing rate. Furtherstudies were conducted all with seeding, at the temperature of around 1°C belowthe freezing temperature of the buffer solution. The freezing rate can then bedefined and controlled precisely.

可见播种的重要性。必须进行播种以达到可控的冷冻速度。进一步的研究都是在接种的情况下进行的，温度比缓冲溶液的冷冻温度低约 1°C。然后可以精确定义和控制冷冻率。

Freezing and Thawing in Potassium Phosphate Buffer 在磷酸钾缓冲液中冻融

Freezingand thawing of protein solutions was investigated, with seeding, using 50 mM potassiumphosphate buffer at pH 7. The effect ofthawing rate on the recovery of LDH activitywas examined first. LDH solutions were frozen at fixed rates of 0.8 and 60°C/min and thawed at different thawing rates from 0.5 to 100°C/min. Theresults are shown in Fig. 3. Theinactivation of LDH after freezing and thaw- ingmarkedly decreased when the thawing rate was in- creased from 0.5 to10°C/min, at which point a plateau was reached.This trend was similar regardless of the difference in freezing rates, although the actual activity recovery was lower when a faster freezing rate was used.This also dem- onstrated that fastcooling (60°C/min) is more damaging thanslow cooling (0.8°C/min) at this protein concentration. The effect of the freezing rate was then investigated for LDH and ADH at two thawing rates: 1°C/min(in the sen- sitive region) and15°C/min (in the plateau region). The patternof freezing rate dependence, as seen in Fig. 4, ex- hibited again a reciprocal effect of freezing and thawing rates, compared to Fig. 3. The dataillustrated that the re- coverydecreased remarkably with the increase of the freez- ing rate and that the maximum recovery occurred with slow freezing (around 1°C/min) and fast thawing.A more severe damage to proteinshappened when fast freezing (>5°C/min) andslow thawing (e.g., 1°C/min) were applied. Microscopic observations showed that the frozen state of the protein solutions was transparent and with large icecrystal textures at slow freezing. Withfast freezing, fine ice crystals were formedand the frozen samples became dim. It was alsonoted that, although the dependence of freezing rate was similar for LDH and ADH, ADH exhibited lesssensitivity to the freezing rate than did LDH.

使用 pH 7 的50 mM 磷酸钾缓冲液对蛋白质溶液的冷冻和解冻进行了研究，并进行了接种。首先检查了解冻速率对 LDH 活性恢复的影响。 LDH 溶液以 0.8 和60°C/min 的固定速率冷冻，并以 0.5 至 100°C/min 的不同解冻速率解冻。结果见图3。当解冻速度从0.5℃/min提高到10℃/min时，冻融后LDH的失活显着降低，此时达到平台期。无论冷冻速度如何，这种趋势都是相似的，尽管使用更快的冷冻速度时实际活动恢复率较低。这也表明，在此蛋白质浓度下，快速冷却 (60°C/min) 比缓慢冷却 (0.8°C/min) 更具破坏性。然后在两种解冻速率下研究了 LDH 和ADH 的冷冻速率的影响：1°C/min（在敏感区域）和 15°C/min（在高原区域）。与图 3 相比，如图 4 所示的冷冻速率依赖性模式再次表现出冷冻和解冻速率的相互影响。速度，并且最大恢复发生在缓慢冷冻（约 1°C/分钟）和快速解冻时。当应用快速冷冻 (>5°C/min) 和缓慢解冻 (例如，1°C/min) 时，对蛋白质的损害更严重。显微镜观察表明，蛋白质溶液的冷冻状态是透明的，在缓慢冷冻时具有大冰晶纹理。随着快速冷冻，形成细小的冰晶，冷冻样品变暗。还注意到，尽管 LDH 和ADH 对冷冻速率的依赖性相似，但 ADH 对冷冻速率的敏感性低于 LDH。

file:///C:/Users/ADMINI~1/AppData/Local/Temp/msohtmlclip1/01/clip_image006.gif

Figure 3. Effectof thawing rate on LDH activity recovery (potassium phosphate buffer, seedingtemperature −1°C in freezing).

file:///C:/Users/ADMINI~1/AppData/Local/Temp/msohtmlclip1/01/clip_image008.gif

file:///C:/Users/ADMINI~1/AppData/Local/Temp/msohtmlclip1/01/clip_image010.gif

Figure 4. Effectof freezing rate on activity recovery of LDH (a) and ADH (b)(potassium phosphate buffer, seeding temperature −1°C, freezing to −30°C).

It is known that denaturation of proteins during freezingis related to the physical and chemical changes of the local environment aroundthe protein molecules during the pro- cess(Franks, 1985). The change of local physical param- eters, such as ice crystal size and morphological microstruc- tures, and chemical parameters, such as solute concentrationand pH, are thought to depend on the freezing rate. Gener- ally, slow freezing causes large,nonuniform ice crystals to beformed. Freeze concentration occurs, in which solutes,possibly including protein molecules, are pushed into non- frozen regions, producing a largeincrease in solute concen- tration and a pH shift due to the buffer saltcrystallization, causing conditions which may lead to dehydration, disso- ciation, or distortion of proteinsand hence inactivation (Franks, 1985). However, fast freezing is believed to pre- vent extensive ice crystal growth andto substantially hinder the freezing concentration and,hence, the possible denatur- ation of protein. However, the present studyshowed an opposite tendency, namely,that at a fixed thawing rate, the high recovery of activity was observed inslow freezing (about 1°C/min) and the fast freezing (>10°C/min) caused moredamage. Potassium phosphate buffer with a low con- centration and neutral pHdoes not change its pH signifi- cantly (<1 Unit) during freezing (van denBerg, 1959), hence pH shift is not the main mechanism for the observed proteindenaturation. The strong correlation of activity re- covery with the well-defined freezing rate suggested that the surface denaturation is the major causeto the protein dam- age during freezing. This confirmed the hypothesis ofStrambini and Gabellieri (1996) and Chang et al. (1996). Protein molecules areadsorbed on the surface of ice crys- tals, which undergo unfoldingand structural change as iden- tifiedby Strambini and Gabellieri (1996). Protein damage is hence directly correlatedto the surface area of the ice crystalsand the texture of ice. In fast freezing (e.g., when the freezing rate > 10°C/min) a large area containingfine ice crystals was formed, which generates a large surface area readilyadsorbing proteins and causing surface dena-turation. At slow freezing rates (e.g., when freezing rate <1°C/min), ice crystals formed were much larger and pro- duced less surface area, resulting in less significantsurface denaturation.

众所周知，冷冻过程中蛋白质的变性与过程中蛋白质分子周围局部环境的物理和化学变化有关（Franks，1985）。局部物理参数的变化，如冰晶大小和形态微观结构，以及化学参数，如溶质浓度和 pH 值，被认为取决于冻结速率。通常，缓慢冷冻会导致形成大的、不均匀的冰晶。发生冷冻浓缩，其中可能包括蛋白质分子的溶质被推入非冷冻区域，由于缓冲盐结晶导致溶质浓度大幅增加和 pH 值变化，导致可能导致脱水、分解的条件- 蛋白质的形成或变形，从而导致失活（Franks，1985）。然而，据信快速冷冻可防止大量冰晶生长并显着阻碍冷冻浓度，从而阻碍蛋白质的可能变性。然而，本研究显示出相反的趋势，即在固定解冻速率下，缓慢冷冻（约1°C / min）和快速冷冻（> 10°C / min）引起的活性恢复率较高。更多的伤害。具有低浓度和中性 pH 值的磷酸钾缓冲液在冷冻期间不会显着改变其 pH 值（<1 单位）（van den Berg，1959），因此 pH 值变化不是观察到的蛋白质变性的主要机制。活性恢复与明确定义的冷冻速率的强相关性表明表面变性是冷冻过程中蛋白质损坏的主要原因。这证实了 Strambini 和Gabellieri (1996) 以及 Chang 等人的假设。 (1996)。蛋白质分子吸附在冰晶的表面，如 Strambini 和Gabellieri (1996) 所确定的那样，它们会发生解折叠和结构变化。因此，蛋白质损伤与冰晶的表面积和冰的质地直接相关。在快速冷冻中（例如，当冷冻速度 > 10°C / min 时）会形成大面积含有细冰晶的区域，这会产生大表面积，容易吸附蛋白质并导致表面变性。在较慢的冷冻速度下（例如，当冷冻速度 < 1°C/min 时），形成的冰晶更大，产生的表面积更小，导致表面变性不太明显。

The dependence on the thawingrate is just the oppositeof the freezing rate (see the trend in Figs. 3 and 4). Slow thawing(<10°C/min) caused more damage to proteins thanfast thawing (>10°C/min), especially significant to the fast frozen samples. In the fast frozensamples, many tiny ice crystalsexist, partly as a result of rapid quenching. Thesystem is not at thermodynamic equilibrium. External dis- turbances such as temperaturevariation would trigger an internalice structure and texture change to reduce the sys- tem energy, typically characterized by simultaneous meltingand growing of small ice crystals into larger ones (Fransen et al.1986)—a process called recrystallization. Recrystalli-zation imposes additional perturbation to protein molecules at the changing ice–solution interface and causesmore dam- age. At slower thawing rate, recrystallization becomes more significant and hence results in moresevere damage to pro- teins. If thethawing rate is fast enough, for example, whenit is greater than 20°C/min (see Fig. 3), the melting rate outweighs the recrystallizationprocess and hence less pro- teindenaturation is expected.

对解冻速率的依赖与冻结速率正好相反（见图 3 和图 4 中的趋势）。慢速解冻 (<10°C/min) 比快速解冻 (>10°C/min) 对蛋白质造成的损害更大，对快速冷冻样品尤其重要。在快速冷冻的样品中，存在许多微小的冰晶，部分原因是快速淬火的结果。系统不处于热力学平衡。温度变化等外部干扰会触发内部冰结构和质地变化以降低系统能量，其典型特征是小冰晶同时融化并生长成大冰晶（Fransen 等人，1986 年）——这一过程称为重结晶。重结晶对不断变化的冰-溶液界面处的蛋白质分子施加了额外的扰动，并导致更多的损害。在较慢的解冻速度下，再结晶变得更加显着，因此对蛋白质造成更严重的损害。如果解冻速度足够快，例如，当它大于 20°C/min（见图 3）时，熔化速度超过再结晶过程，因此预计较少的蛋白质变性。

On the other hand, recrystallization during thawing ofslow frozen samples may play a less important role. This is because athermodynamic equilibrium is largely maintained during freezing within thesystem. The enzyme activity is hence less sensitive to thawing rate (see Fig.3).

另一方面，慢速冷冻样品解冻过程中的重结晶可能起不太重要的作用。这是因为在系统内冻结期间很大程度上保持了热力学平衡。因此酶活性对解冻速率不太敏感（见图 3）。

Freezing and Thawing in Different Buffer Solutions 在不同的缓冲溶液中冻融

Theeffect of different buffer solutions on freezing damage of LDH was examinedwith sodium phosphate, Tris-HCl,

HEPES,and citrate buffers (0.05 M, pH 7).Samples were seeded at −1°C and frozen at the rates of 0.75 and 30°C/min with thawing rates of 1 and15°C/min, respectively. The activityrecovery of LDH at different buffers after freezing and thawing was showed inFig. 5. Compared with Figs. 3 and 4, the dependence on the freezing and thawingrates for all the buffers follows the same pattern. As seen in Fig. 5, the buffer did affect activityrecovery, and sodium phos- phatebuffer gave the lowest recovery at given freezing and thawing protocols.

用磷酸钠、Tris-HCl、HEPES和柠檬酸盐缓冲液 (0.05 M, pH 7) 检查不同缓冲溶液对 LDH 冷冻损伤的影响。样品在 -1°C 下接种，并以 0.75 和30°C/min 的速率冷冻，解冻速率分别为 1 和15°C/min。冻融后不同缓冲液中LDH的活性恢复见图5。如图 3 和 4 所示，所有缓冲液对冷冻和解冻速率的依赖遵循相同的模式。如图 5 所示，缓冲液确实影响了活性恢复，磷酸钠缓冲液在给定的冷冻和解冻方案下的恢复率最低。

The significant pH shift of sodium phosphate buffer, due to the precipitation of the buffer saltduring freezing, may contribute to the freeze denaturation of LDH, as LDH is pH sensitive assembly (Anchordoquy andCarpenter, 1996). It was observedthat the pH of sodium phosphate solution candrop more than 3 units during freezing. In comparison, Tris,citrate, and potassiumphosphate buffers change pH by +1, −0.4, and +0.6 during freezing (Jiangand Nail, 1998; Larsen, 1973; Orii and Morita, 1977; van den Berg and Rose1959). It is therefore not surprising to see LDH in sodium phosphate suffersthe most damage buffer during freezing.

由于冷冻过程中缓冲盐的沉淀，磷酸钠缓冲液的显着 pH 变化可能有助于 LDH 的冷冻变性，因为 LDH 是pH 敏感组件（Anchordoquy 和 Carpenter，1996）。据观察，磷酸钠溶液的 pH 值在冷冻过程中会下降超过 3 个单位。相比之下，Tris、柠檬酸盐和磷酸钾缓冲液在冷冻过程中会改变 pH 值 +1、-0.4 和 +0.6（Jiang 和 Nail，1998 年；Larsen，1973 年；Orii 和 Morita，1977 年；van den Berg 和 Rose 1959 年）。因此，看到磷酸钠中的 LDH 在冷冻过程中受到最大的损害也就不足为奇了。

Thiseffect was further demonstrated by the addition of NaCl into potassium phosphate buffer. Potassium phosphate and sodium chloride form Na2HPO4,which can precipitate at low temperature and then changethe pH of buffer dra- matically(Chilson et al., 1965; Murase and Franks, 1989). The experimental results areshown in Figs. 6 and 7, which give similar patterns of dependence of activityrecovery on freezing and thawing rates to those in sodium phosphate buffer.Freezing damage increased with addition of salt in the whole range of thefreezing rate studied, similar to the observation of Fishbein and Winkert(1977) in their experi- mental work with catalase. It is quite possible thatthe in- crease of the concentration of sodium chlorideleads to more crystallizationof Na2HPO4 duringfreezing and therefore a further pHdrop in the unfrozen buffer (Larsen, 1973). Thedecreased activity recovery at given rates of freezing and thawing may be attributed to theacidification of the buffer solutions when the concentration of sodium chlorideis in- creased (see Figs. 6 and 7).

通过将 NaCl 添加到磷酸钾缓冲液中进一步证明了这种效果。磷酸钾和氯化钠形成 Na2HPO4，Na2HPO4 可以在低温下沉淀，然后显着改变缓冲液的 pH 值（Chilson 等，1965；Murase 和Franks，1989）。实验结果如图 1 和图 5 所示。图 6 和7 给出了与在磷酸钠缓冲液中的那些相似的活性恢复对冷冻和解冻速率的依赖性模式。在研究的整个冷冻速率范围内，冷冻损伤随着盐的加入而增加，类似于 Fishbein 和Winkert (1977) 在他们使用过氧化氢酶的实验工作中的观察结果。很有可能氯化钠浓度的增加导致冷冻过程中更多的 Na2HPO4 结晶，从而导致未冷冻缓冲液中的 pH 进一步下降（Larsen，1973）。在给定的冷冻和解冻速率下活性恢复的降低可能是由于氯化钠浓度增加时缓冲溶液的酸化（见图 6 和7）。

file:///C:/Users/ADMINI~1/AppData/Local/Temp/msohtmlclip1/01/clip_image012.jpg

Figure 5. Effect of buffertypes on LDH activity recovery at thawing rate of 15°C/min (A) and 1°C/min (B).

file:///C:/Users/ADMINI~1/AppData/Local/Temp/msohtmlclip1/01/clip_image014.gif

Figure6. Effect of freezing rate on activity recovery inpotassium phos- phate solution with added NaCl (a, LDH thawing rate 1°C/min; b,ADH thawing rate 15°C/min).

file:///C:/Users/ADMINI~1/AppData/Local/Temp/msohtmlclip1/01/clip_image016.gif

Figure 7. Effect of NaCladdition on the activity recovery of catalase (thawing rate 1°C/min).

From Fig. 6a it can be seen that an optimal freezing rateseems to exist around 1°C/min when NaCl is added into the potassium buffer.When the freezing rate is lower that this, protein damage is increased as thefreezing rate decreases. This phenomenon is not observedwith potassium phosphate without NaCl addition, thereforeit should be related to the increased salt precipitation and pH shift when thefreezing rate is below 1°C/min. It is likely that at such a low freezing rate, a chemical and thermodynamicequilibrium is achieved,leading to maximum salt precipitation and pH change. Such an optimal freezingrate may exist for ADH but it is not so obvious (see Fig. 6b), and it is notobserved for catalase (Fig. 7).

从图 6a 可以看出，当将 NaCl 添加到钾缓冲液中时，最佳冷冻速率似乎存在于 1°C/min 左右。当冷冻速度低于此值时，蛋白质损伤会随着冷冻速度的降低而增加。未添加 NaCl 的磷酸钾未观察到这种现象，因此当冷冻速度低于 1°C/min 时，这应该与盐析出增加和 pH 变化有关。很可能在如此低的冷冻速率下，实现了化学和热力学平衡，导致最大的盐沉淀和 pH 值变化。对于 ADH 可能存在这样的最佳冷冻速率，但它并不那么明显（见图 6b），并且在过氧化氢酶中没有观察到（图 7）。

Based on this work, it could be concluded that proteindenaturation during freezing, in the absence of protecting chemicals, is mainlycaused by surface denaturation and relatedto exposure to the ice–water interface. It isadvisable to use a relatively slow freezing (about 1°C/min) and fast thawing process (>10°C/min). Sodiumphosphate buffer and othersolutions with a tendency of salt precipitation during freezing should be avoided. An optimal freezing rate mayexist that depends on the solution conditions.

基于这项工作，可以得出结论，在没有保护化学品的情况下，冷冻过程中的蛋白质变性主要是由表面变性引起的，并且与暴露于冰水界面有关。建议使用相对缓慢的冷冻（约 1°C/min）和快速解冻过程（>10°C/min）。应避免在冷冻过程中使用磷酸钠缓冲液和其他有盐沉淀倾向的溶液。可能存在取决于溶液条件的最佳冷冻速率。

It should be pointed out that the above conclusions are drawn from the experimental data usingof a dilute protein concentration(0.025 g/l), the effect of protein concentrationshould be studied in future. Also the effect of salt type and cryoprotecting chemicals on freezingdenaturation of pro- teins should bequantified in future research.

需要指出的是，上述结论是从使用稀释蛋白质浓度（0.025 g/l）的实验数据得出的，蛋白质浓度的影响有待进一步研究。此外，盐类和冷冻保护化学品对蛋白质冷冻变性的影响应在未来的研究中进行量化。

References 参考

AnchordoquyTJ, Carpenter JF. 1996. Polymer protect lactate dehydroge- nase duringfreezing–drying by inhibiting dissociation in the frozen state. Arch BiochemBiophys 332:231–238.

ArakawaT, Prestrelsk SJ, Kenney WC, Carpenter JF. 1993. Factors af-

fectingshort-term and long-term stabilities of proteins. Adv Drug Del Rev 10:1–28.

Chang BS, Kendrick BS, Carpenter JF. 1996. Surface-induced denaturation of proteins during freezing and its inhibition by surfactant. J Pharm Sci 85:1325–1330.

ChilsonOP, Costello LA, Kaplan NO. 1965. Biochemistry 4:271–281. Eckhardt BM, OesweinJQ, Bewley TA. 1991. Effect of freezing on ag-

gregation of human growth hormone. Pharm Res 8:1360–1364.

FishbeinWN, Winkert JW. 1978. Parameters of biological freezing dam- age in simplesolutions: catalase. II. Demonstration of an optimum- recovery cooling-ratecurve in a membraneless system. Cryobiology 15:168–177.

FranksF. 1985. Biophysics and biochemistry at low temperature. London: CambridgeUniversity Press.

FransenGJ, Salemilk PJ, Cormmelin DJA. 1986. Critical parameters in freezing ofliposomes. Int J Pharm 33:27–35.

IzutsuK, Yoshioka S, Terao T. 1994. Stabilizing effect of amphiphilic excipients onthe freezing–thawing and freezing–drying of lactate de- hydrogenase. BiotechnolBioeng 43:1102–1107.

JiangS, Nail SL. 1998. Effect of process conditions on recovery of protein activityafter freezing and freezing–drying. Eur J Pharm Biopharm 45:249–257.

LarsenSS. 1973. Studies on stability of drugs in frozen systems. VI. The effect of freezing upon pH for buffered aqueoussolutions. Arch Pharm ChemSci Ed 1:41–53.

MuraseN, Franks F. 1989. Salt precipitation during the freeze–concentra- tion ofphosphate buffer solutions. Biophys Chem 34:293–230.