97

fOCUS

Why are enzymes less active in

organic solvents than in water?
Alexander M. Klibanov

In order to exploit fully the biotechnological opportunities afforded by nonaqueous

enzymology, the issue of often drastically diminished enzymatic activity in organic
solvents compared with that in water must be addressed and resolved. Recent
studies have made great strides towards elucidating causes of this phenomenon
of activity loss. None of these causes is insurmountable; by designing strategies
that systematically target them, enzymatic activity in organic solvents can be
readily enhanced by multiple orders of magnitude and ultimately brought to the
aqueous-like level.

Over the past decade, nonaqueous enzymology has example, in the case ofot-chymotrypsin and subtilisin,
emerged as a major area ofbiotechnology research and the specificity constant, kcat/KM, of an enzymatic trans-
development ~. Although the notion of enzymatic esterification reaction in anhydrous octane exceeds the
processes in organic solvents was initially met with bimolecular rate constant of the same reaction with-
scepticism by the scientific community, nonaqueous out enzyme by as nmch as 1011-fold 4. Second, whereas
enzymology is now studied and exploited by numer- the rates of enzymatic processes in anhydrous solvents
ous academic and industrial laboratories worldwide, are indeed dwarfed by those in water, this is not
and the number of papers devoted to this topic has necessarily so in hydrous ones. For example, poly-
surged into the hundreds 1. phenol oxidase in octanol containing 3% water retains
The reason why nonaqueous enzymology has gen- more than a third of its activity" in water s. Likewise,
era~:ed so nmch excitement is that enzymes exhibit the activity, of yeast alcohol dehydrogenase in isopropyl
striking new properties in organic solvents. For ex- ether containing 0.5% water is about a quarter of that
ample, in organic solvents enzymes often catalyze reac- in aqueous solution s .
tions impossible in water and are very stable 1. Further- These comments notwithstanding, the fact that
more, enzyme selectivity in organic solvents is not only enzymatic activity, in anhydrous or nearly anhydrous
distinct from that in water but can be markedly con- solvents falls short of that in water clearly threatens
trollied, and even reversed, by the solvent2.3; this the practical potential of nonaqueous enzymology
'soNent engineering' thus provides an alternative to (although it should be kept in mind that from the
protein engineering. industrial standpoint the real question is whether the
A number of applications for enzymes in organic enzymatic activity" is high enough for the intended
solvents have been developed in chemical processing application, not how it compares with that in water).
(particularly for the synthesis of optically active inter- Therefore, the issue of activity loss must be confronted
mediates), food-related conversions and analyses 1. The and, hopefully, alleviated. The only way to do so ration-
ordy dark cloud hanging over nonaqueous enzymol- ally is to elucidate why enzymatic activity plummets
og2,' is the notion (largely, albeit not always, correct) on transition from water to anhydrous solvents. This
that enzymes in organic solvents are far less active than elucidation, much advanced by recent studies, is the
in water. For instance, the proteases ot-chymotrypsin subject of the present article. The discussion will
and subtilisin have activities 104-105-times lower in involve, primarily; the model enzyn]e subtilisin Carls-
anhydrous octane than in water; the two enzymes are berg, for which the bulk of the relevant work has been
less active still in most other organic solvents4. Such done and the greatest insights have been attained. Two
numbers seem to be typical, although two comments enzyme forms introduced into organic solvents will be
must be made. First, while handicapped by transfer considered: amorphous (e.g. lyophilized, which is
from water to organic solvents, enzymes nevertheless the most popular form among researchers 1 s) and
remain impressive catalysts in organic solvents. For crosslinked crystalline (CLC) enzymes 6.

A. "~[Link],mov (klibanov@[Link]) is at the DctJartment of Chem- Diffusion and accessibility factors

istr]', Massachusetts Institute qf Tectmolog),, (~amtm'd.~e, MA 02139, Enzymes are soluble in water but insoluble in nearly
US d. all organic solvents. Hence, the addition of enzymes

Cop?right © 1997, Elsevier Science Ltd. All rights reserved. 0167 - 7799/97/S17.00. PII: S0167-7799(97)01013-5 TIBTECH MARCH 1997 {VOL 15)
 98

fOCUS
leads to solutions in w'ater but, usuall?, suspensions in must underlie, for example, the 10V-fold difference
organic solvents. Consequently, it is intuitively plaus- in enzymatic activity. (k~a~/K,a) between subtilisin
ible that enzymatic activity in nonaqueous solvents dissolved in water and CLC subtilisin suspended in
may be diminished due to diffusional limitations on acetonitrile~ ~.
the substrates. This phenomenon, common in hetero- X-ray crystallography, unfortunately, cannot be
geneous, including immobilized enzyme, catalysis7, applied to lyophilized enzymes suspended in organic
results in the under-utilization of enzymatic power and solvents, thus making complete structure characteriz-
thus decreased observed activity. Theoretical analysis s ation of such systems impossible. Nevertheless, some
shows that such mass transfer limitations may indeed structural information can be derived by other bio-
slow down enzymatic catalysis in organic solvents. We physical methods, such as Fourier-transform infrared
experimentally tested this possibility in the specific (FTIR) spectroscopy. Using this methodology we have
case of subtilisin. If subtilisin catalysis in organic sol- recently ascertained that placing lyophilized subtilisin
vents is suppressed by diffusional limitations, then the in a variety" of anhydrous solvents, such as octane,
dependence of the enzymatic activity in such media, acetonitrile and dioxane, has no appreciable effect on
or that in water, should be nonlinear (and should its secondary, structure (as reflected by the oL-helix con-
eventually level off-); by contrast, a linear dependence tent) 16. However, the preceding step, lyophilization,
should ensue in the absence of diffusional restric- does result in a significant (and reversible iflyophiliz-
tions 7,9. In fact, we observed linear dependences in all ation is done carefully) denaturation ofsubtilisin 16 and
three organic solvents examined: dodecane, carbon many other proteins 17.1~. In other words, somewhat
tetrachloride and acetonitrile 9. Therefore, the 5t)00- ironically, it is not the contact with an organic sol-
100000 times slower catalysis in these solvents con> vent but the prior dehydration that alters the enzyme
pared with that in water 4 cannot be caused by mass structure and, undoubtedly, lowers catalytic activity.
transfer limitation. This deleterious effect can be prevented, or at least
Even in the absence ofdiffusional limitations per se, minimized, by lyoprotectants, as discussed in a later
it is conceivable that in a lyophilized enzyme particle section.
or CLC enzyme, some enzymatic active centers may
be sterically shielded by the neighboring enzyme mol- Energetics of substrate desolvation and transition
ecules and thus be inaccessible to the substrate. This state stabilization
would effectively prevent those enzyme molecules The energy, of binding betw'een the enzyme and the
from participating in catalysis, thereby cutting the substrate is the major driving force for enzymatic cataly-
observed catalytic activity: This possibility was tested sis and the source o f huge rate enhancements afforded
by titrating the competent active centers in subtilisin by enzymes ~9. In order for binding to occur, the sub-
and c(-chymotrypsin suspended in organic solvents. strate must first undergo desolvation upon its journey
The data obtained<> reveal that in octane the nun> from the reaction medium to the enzyme active cen-
bet of centers is between one- and two-thirds of that ter. The more energetically favorable this desolvation
in water. For CLC subtilisin in organic solvents it is is, the greater the net binding energy (resulting in
even higher - essentially unity 11. Therefore, steric catalysis) becomes. Subtilisin, (x-chymotrypsin and
blockage can be responsible for only a tiny fraction of many other enzymes have hydrophobic active centers;
the loss of enzymatic activity on transition ti'om water they work best with hydrophobic substrates because
to organic solvents. there is a large energetic incentive for the latter to par-
tition from water into the active center. When water
Structural changes is replaced with an organic solvent, the substrate is no
Another obvious suspect among the possible causes longer 'squeezed out' of the medium owing to the
of the diminished enzymatic activity in anhydrous sol- hydrophobic effect and the energetic advantages of the
vents compared with that in water is a change in partitioning drop - or, in thermodynamic language,
enzyme conformation. Such protein denaturation m the ground state of a hydrophobic substrate is stabi-
aqueous-organic mixtures is well known 12, and thus lized in organic solvents relative to water 2°. This ought
it would not be unreasonable to expect an even graver to raise the activation barrier and hence slow down the
situation in neat organic solvents. For CLC enzymes enzymatic reaction > .
suspended in anhydrous solvents, this issue can be Using CLC subtilisin (where, as discussed above,
addressed experimentally by means of X-ray crystal- there are no conformational changes upon replace-
lography. Several years ago we succeeded in solving ment of water with anhydrous reaction media), we
the crystal structure of CLC subtilisin in anhydrous have quantified the decline in enzymatic activity
acetonitrile 13. Comparison of this structure with its stemming from this less favorable substrate desolvation.
counterpart in water L~,~4revealed that the two are vir- In the case of the substrate N-Ac-k-Phe-OEt, this
tually identical. The same holds true for the recently effect accounts for more than a 100-fold reduction in
solved crystal structure of CLC subtilisin in dioxane k , U K M in anhydrous acetonitrile compared with that
(see the cover of this journal issue). Likewise, the X- in water I i.
ray crystal structure of uncrosslinked chymotrypsin in Another term aftbcting the activation energy (and
hexane was found to be very similar to that in water ~5. thus enzymatic reactivity) is the energy of the transi-
Therefore, factors other than conformational change tion state. For subtilisin and many other hydrolases, this

TIB1-ECH MARCH 1997 (VOL 1.5)

 99

fOgPlS
reaction transition state is highly polar (resembling a they strongly bind some 'essential' water and retain it
charged tetrahedral intermediate) 19. If it is fully even when suspended in anhydrous solvents 4.
enveloped by the enzyme active center (i.e. shielded There are numerous lines of evidence supporting
from the solvent), a solvent change should not affect this analysis. There is a general inverse correlation
its energy. However, full shielding does not always between enzymatic activity in an organic solvent and
o c c u r - for example, at least one-third of the transition the latter's hydrophilicity - the highest activity is
state ofsubtilisin with N-Ac-L-Phe-OEt is exposed to usually observed in hydrophobic solvents 4,25. This is
the solvent 2. Since water stabilizes tetrahedral inter- because hydrophobic solvents possess a lesser ability
mediates nmch better than less polar solvents, the than their hydrophilic counterparts to strip the essen-
enzymatic activity in the latter should be lower as a tial water offenzyme molecules 4. Addition of water to
result of this phenomenon 21. Enzyme-substrate inter- enzyme suspensions in anhydrous solvents s or increas-
actions in partially-exposed transition states may also ing the thermodynamic activity of water (aw) in such
be affected by the solvent > . systems by other means 2(' can dramatically enhance
enzymatic activity. To some extent, water can be
Conformational mobility replaced by water-mimicking organic solvents, such as
I:ateraction of proteins with water (hydration) is glycerol or ethylene glycol, which can also form mul-
critical for a plethora of biological functions, particu- tiple hydrogen bonds 5.2:.
larly enzymatic activity, 2-~. Water, acting as a lubricant Note that CLC subtilisin loses almost two orders of
or plasticizer23, allows enzymes to exhibit the confor- magnitude in catalytic activity solely due to the decline
madonal mobility required for optimal catalysis. in its conformational mobility in anhydrous aceto-
Organic solvents, in general, are not nearly as accom- nitrile compared with that in water ~.
modating, if accommodating at all, because they lack
wal:er's ability to engage in nmltiple hydrogen bonds pH situation
(and, because of their lower dielectric constants, lead In aqueous solution, enzyme catalysis profoundly
to ~;tronger electrostatic interactions and hence more depends on the pH and has a pH optimung'. Since
r i g d proteinse4). These considerations dictate that, the notion o f p H has no meaning in organic solvents,
other things being equal, enzymes should be less active the protonation state of enzyme ionogenic groups
in ~mhydrous solvents than in water owing to restricted must be controlled by other factors. It was discovered4,2s
conformational mobility. In fact, enzymes exhibit that enzymes have a 'pH memory' - i.e. their catalytic
significant catalytic activity in such media only because behavior in organic solvents reflects the pH of the last

Fable 1. Causes of lower enzymatic activity in organic solvents compared with that in water and ways to
alleviate them a

Cause b Comments Remedies b

E~iffusional limitations Not as likely as widely claimed Vigorously agitate enzyme suspensions;
use small enzyme particles
Active center blockage Responsible for no more than a few-fold If matters, use crystalline rather than
activity reduction amorphous enzyme particles
C:onformational change Occurs upon lyophilization and other Use lyoprotectants; alternatively, prepare
modes of dehydration. Is avoided with enzyme complexes with amphiphiles
CLC enzymes. Is usually not caused by soluble in organic solvents;
contact with the solvent use enzyme CLCs
Unfavorable energetics Severe with hydrophobic substrates; Select the solvent expected to yield
of substrate desolvation also, likely with natural substrates. May unfavorable solvent-substrate
be responsible for at least a lO0-fold interactions
activity reduction
-[ransition state Likely to be pronounced when the Select the solvent expected to yield
destabilization transition state is at least partially favorable interactions with the transition
exposed to the solvent state
Reduced conformational Severe in anhydrous, hydrophilic solvents Optimize water activity (aw); hydrate the
mobility owing to the stripping of the essential solvent; use hydrophobic solvents; use
enzyme-bound water. May be responsible water-mimicking and denaturing
for at least a lO0-fold activity reduction cosolvent additives
Suboptimal pH situation May be responsible for at least a lO0-fold Dehydrate from aqueous solution of the pH
activity reduction optimal for enzymatic activity; use
organic-phase buffers

a'Fhetable refers to lyophilized(or other amorphouspreparations)or crosslinkedcrystalline(CLC)enzymessuspendedin neat organic

salvents.
bSoth proven and hypothetical(see text).

TIBTECHMARCH1997(VOL15)
 100

fOCUS

aqueous solution to which they were exposed (e.g. in water. In fact, it may even be possible to make
from which they" were lyophilized). Therefore, unless enzymes more active in organic solvents than in water.
enzyme powders are obtained from an aqueous solu- For example, particularly for unnatural substrates, the
tion of the pH affording maximal activity, enzymatic energetics ofsubstrate desolvation may be made more
performance in organic solvents is doomed to be favorable. Also, there is the intriguing possibility of
suboptimal. molding enzyme active centers and thus optimizing
An alternative to recovering enzymes from aqueous enzymatic activity in organic solvents for a desired
solutions at appropriate pH values is to employ substrate by molecular imprinting32.37.
organic-phase buffers - i.e. suitable mixtures of There is now convincing evidence dispelling the
organic-soluble acids and their conjugate bases 2s. This popular clich6 that enzymes should be more active in
approach has been successfully used to increase water than in organic solvents because that is where
markedly the catalytic activity of lyophilized-'8 and Nature intended to use them. Regardless of Nature's
CLC (1Kef. 29) enzymes in organic solvents. intention, if optimally used, enzymes may be able to
work as well in nonaqueous as in aqueous media. The
How to enhance enzymatic activity in organic real question thus becomes: what novel mechanistic
solvents knowledge and useful practical applications will be
The foregoing analysis shows that there are several derived from this opportunity?
physicochemical factors that may lower enzymatic
activity in organic solvents compared with that in Acknowledgements
water (Table 1). Gratifyingly, virtually all examples of Work in the author's laboratory was supported by
enzyme activation in nonaqueous media reported in grants from the National Institutes of Health
the literature to date can be explained in ternxs of (GM39794) and Department of Energy (BCTR.
minimizing these individual factors. program). I am grateful to Jennifer L. Schmitke for
The most popular strategy has been to prevent helpful discussions.
enzyme denaturation during dehydration. This has
been achieved by lyophilizing enzymes in the presence References
of lyoprotectants such as sugars 3°, poly(ethylene gly- 1 Koskinen, A. M. P. and Klibanov, A. M., eds (1996) EnzymaticReac-
col) 3°, inorganic salts (notably KC1)31, substrate-resem- tions i, Organic Media, Blackie Academic & Professional
bling ligands>.32, and crown ethers 33. An alternative is 2 Wescott, C. IX. and Klibanov, A. M. (1994) Biochim. Bioph U, Acta
to preform organic-soluble complexes of enzymes 1206, 1-9
3 Carrea, G., Ottolina, G. and Riva, S. (1995) Teen& Bioteehnol. 13, 63-70
with detergent (or other amphiphile) molecules3<35; 4 Zaks, A. and Klibanov. A. M. (1988)]. Biol. Chem. 263, 3194-3201
enzymes are apparently in native-like conformations 5 Zaks, A. and Klibanov, A. M. (1988)_I. Biol. Chem. 263, 8017-8021
in such complexes 16,3s. These approaches have resulted 6 Margolin, A. L. (1996) Trends Biotedmol. 14, 223-230
in increasing enzymatic activities in organic solvents 7 Boudart, M. and Burwell, IZ. L. (1973) in lm,estl~ation of Rates and
by up to three to four orders of magnitude. Mecilanisms qfReaction~ (Lewis, E. S., ed.), pp. 693-739, Wiley
Another effective strategy has been to 'loosen up' 8 Kamat, S., Beckman. E.J. and Russell, A.J. (1992) Enzyme Microb.
Tedmot. 14, 265-271
enzymes in anhydrous solvents. Enzymatic activity has
9 Wescott, C. K. and Klibanov. A. M. (1993)_J. Am. Chem. Soc. 115,
been increased up to two to three orders of magnitude 1629 1631
by adding small quantities of water to organic sol- 10 Wangikar, P. P.. Carmichael, 1)., Clark, D. S. and Dordick, J. s.
vents 5, maximizing a,,, (P,,ef. 26), and adding water (1996) Biotechnot. Bioen~q. 511, 329-.335
mimics 5,27 or denaturing cosolvents 3~' to anhydrous 11 Schmitke, J. L., Wescott, C. R. and Klibanov, A. M. (1996)J. Am.
media. Chem. S0c. 118, 336{>3365
12 Lapanje, S. (1978) Phl'sicodtemical Aspects of Protein Denaturation,
Ch. 3, 4 and 6, Wiley
Concluding remarks 13 Fitzpatrick, P. A.. Steimnetz, A. C. U., Ringe, D. and
The reasons why enzymatic activity in organic sol- Klibanov, A. M. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 8653-8657
vents usually does not measure up to that in water have 14 Fitzpatrick, P. A., Kinge, I). and Klibanov, A. M. (1994) Biochem.
been elucidated (Table 1). In some instances, their Bioph U. Res. Commtm. 198, 675-681
contributions to the overall activity loss have been 15 Yetmawar, N. H.. Yennawar, H. P. and Farber, G. K. (1994)
quantified, and the sum thereof comes close to the Biochemistry 33, 7326 7336
16 Gnebenow, K. and Klibanov, A. M. (1997) Biotedlnol. Bioen~.
observed decline in the catalytic performance in non-
54 (in press)
aqueous media compared with water u. None of these 17 I)ong, A., Prestrelski. S.J., Allison, S. D. and Carpenter, J. F. (1995)
factors is insurmountable and, as the aforementioned J. Phaml. Sci. 84. 415-424
examples illustrate, can be prioritized, systematically 18 Griebenow, K and Klibanov, A. M. (1995) Proc. Natl. ArM. &i.
tackled (Table 1) and, perhaps, eliminated. Indeed, l]. S. A. 92, 11)969-10976
'aqueous-like' activities of enzymes in nonaqueous 19 Fersht, A. (1985) Ett~.ymc Structure arid Mechanism (2nd edn),
solvents have thus been achieved5,3s and less spectacu- Freelnan
20 Wan~kar, P. P., Kich, J. O , Clark, D. S. and Dordick, J. s. (1995)
lar but nonetheless quite dramatic activations are now
Biochemistry 34, 123(12-1231{I
commonplace 4,>,1~,26-36. There is every reason to be 21 Xu. Z-F., Affieck, K., Wanokar, P., Suzawa, V., Dordick, J. S. and
optimistic that, by judiciously combining the rem- Clark, 1). S. (1994) Biot~vtmot. Bioe~g. 43, 515-520
edies discussed above, enzymatic activities in organic 22 Rupley, J. A. and Caren, G. (1991) Adv. Protein Chem. 41,
media can be routinely made comparable with those 37-172

TI[ffECH MARCH 1997 (VOL ] 5)

 101

23 (;regory, R. B. (1995) in Protein-Solvem lnteractiom (Grego~', R. B., 31 Khmdnitsky, Y. L., Welch, S. H., Clark, D. S. and Dordick, J. S.
ed.), pp. 191-264, Marcel Dekker (1994)J. A m Chcm Soc. 116, 2647-2648
24 Pal:fleck,R., Haynes, C. A. and Clark, D. S. (1992) Proc..N'atl. AcM. 32 Russell, A. J. and Klibanov, A. M. (1987).I. Biol. Chem. 263,
Sci. (r. S. A. 89, 5167 517(I 11624 11626
25 Zaks, A. and Klibanov, A. M. (1985) Pro, ,M~tl. Acad. Sci. U. S. A. 33 Broos, J.. Sakodinskaya, 1. K., Engbersen, J. F. J., Verboon, W. and
82, 3192-3196 [Link], D. N. (1995).1. Chem. &,c., Chem. Commlm. 255-256
26 Bell, G., Hailing, P.J., Moore, B. 1).. Partndge. J. and Rees, 1). G. 34 Okahata, Y. and ljiro, K. (1992) Bull. Chef,,. Soc. Jpn. 65,
(1995) Trotds Biotechnol. 13, 468-473 2411-2420
27 Kitaguchi, H. and Klibanov, A. M. (1989) j. Am. (3win. S0c 1 I1. 35 Paradkar, V. M. and Dordick, J. S. (1994)J. Am. Chem. Soc. 116,
c~272-9273 5009-5111()
28 [;lackwood, A. D., Curran. L. J., Moore, B. 1). and Hailing, P. J. 36 Ahnarsson, O. and Klibanov, A. M. (1996) Bio.'dmoL Bioen,q. 49,
(1994) Biochim. Biophys. Acta 1206, 161-165 87-92
29 )[u, K. and Klibanov, A. M. (1996)J. Am. (ghent. Soc. 118. 9815 9819 37 Mingan-o, i.. Gonz~ilez-Navarro, H. and Braco, L. (1996) Biochemistry
30 I)abulis, K. and Klibanov, A. M. (1993) Biotedmot. Bioe,(e. 41. 566-571 35, 9935-9944

Atomic force microscopy in

analytical biotechnology
Stephanie Allen, Martyn C. Davies, Clive J. Roberts, Saul J. B. Tendler and Philip M. Williams

The past decade has seen the atomic force microscope evolve not only as a high
resolution imaging tool, but also as an instrument capable of measuring forces
between surfaces and the material properties of samples. Here, the use of atomic
force microscopy for surface force measurement is reviewed, highlighting the
considerable progress that has recently been made in the area of biotechnology.
Particular emphasis is placed on how the instrument can be used to probe directly
biomolecular interactions, illustrating the potential of the technique to impact on
our fundamental understanding of molecular structure and processes.

Since its invention in 1986, the atomic force nficro- T h e A F M as a force s e n s i n g i n s t r u m e n t

scope (AFM) 1 has evolved as a valuable imaging Intermolecular forces govern the fundamental prop-
tec]mique with resolution in the micrometre to sub- ertics of solids, liquids and gases, the properties o f col-
nanometre range. The ability of the AFM to inaage loids and the organization of biological structures s.
both insulating and conducting surfaces in a varie D" of Nmnerous biophysical methods have been employed
environments has facilitated molecular resolution to investigate specific and non-specific molecular
images of a wide range ofbiomolecules, including pro- interactions, including optical trapping 9, magnetic
teir:.se, lipids 3 and D N A ([Link]. 4). In addition to its force experiments >, pipette suction1' and the surface
imaging abilities, the AFM can also be e m p l w e d to force apparatus~L Atomic force nficroscopy comple-
probe spatial variations in surface properties, such as ments these techniques by uniquely combining an
adhesiveness and elasticity 5. The use of the AFM as an extended force range with high spatial resolution. The
imaging technique in the study of biomolecules has AFM has a theoretical force sensitivity of 10 1~N (Ref.
frec:uently been the subject of review <7. The aim of 13), and employs probes with a tip apex of 10-50 nm
this article is to outline the use of an AFM as an instru- radius 14, to produce contact areas as small as 10 nm2;
ment for surface force measurements, with a particu- therefore providing a means to overcome the limi-
lar emphasis on how the technique can be employed tations of other techniques.
in the areas ofbiotechnolog3/and biological science. By recording force measurements between different
probe-sample combinations in various ionic media, it
was first illustrated that the AFM could be used to
S. ~:lle~ (PAXSA@[Link]), 31. C. Davies, C.J. Roberts,
S. J B. Temtler a~d P. M. HTlliams are at the L.,lbomtor}, of Bio-
measure forces, such as van der Waals and electrosta-
plly~ics and Surface Analysis, Department of l)harmaceutical Sciemcs, tic forces 15. Ducker and colleagues 1(~also employed the
The Universit), qf NottiuA~ham, [hlil,ersit), Park, .\\ntinA~hanl, U K technique to measure colloidal forces, by attaching
N G 7 2RD. spherical particles to the apex of AFM cantilevers. The

Copyright © 1997, Elsevier Science Ltd. All rights reserved. 0167 7799/97/$17.06. PII: S0167-7799(97)01015-9 TIBTECH MARCH 1997 IVOL 15)