Understanding enzyme action at solid surfaces
Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow G1 1XL, Scotland, U.K.
In solid-phase synthesis, there is interest in using enzymes that normally act on dissolved substrates. It is
normally observed that rates and yields are substantially reduced when the usual substrates are covalently
attached to a solid particle. Recently, there has been some progress in understanding the reasons for this,
and hence how to improve behaviour. Diffusion of enzyme molecules into some of the support particles
used in solid-phase chemistry is slow or absent. Methods are now available to visualize the sites of reaction,
and hence detect this problem, and identify better support materials. Chemical equilibrium positions for
reactions at the surface can be substantially altered compared with those in solution, so may unexpectedly
limit yields. The shift can also be exploited to carry out, for example, direct synthesis of peptide bonds in
an aqueous environment. The rate of enzyme attack depends on how the substrate moiety is attached to
the surface, with an optimal ‘spacer’ length.
Introduction systems, for example when used to assay activity. The present
Most enzyme-catalysed reactions involve substrates that are paper will deal with three questions that have been addressed
dissolved in the reaction medium. Some enzymes, such as in work at Strathclyde (in collaboration with others). Can
cellulases and amylases, naturally act at the surface of insol- the enzyme reach the sites of reaction in a porous support
uble substrates, and have special features to facilitate this. But particle? How will the equilibrium position of the reaction
there is increasing interest in enzymes that normally have be affected? What is the influence of the length of a ‘spacer’
dissolved substrates acting instead on groups attached to solid attaching the substrate moiety to the surface?
supports. These systems can be seen as complementary to the
more familiar immobilized enzymes and indeed have been
referred to as immobilized substrates. Accessibility of enzyme to sites
Enzyme action on surface-attached substrates is relevant in within beads
a number of situations. In solid-phase synthesis, the specifi- An obvious first condition for catalysis is that the enzyme
city of enzymes can be valuable [1,2], particularly for re- molecules reach the sites at which substrate moieties are
moving or avoiding the need for protecting groups, or attached. In solid-phase chemistry, there is a natural pref-
selectively cleaving linkers in a final step. Where combi- erence for support materials with a high density of sites for
natorial libraries of potential enzyme substrates are prepared attachment of the species being synthesized. This is achieved
on solid supports, it is convenient to be able to assay them by using support materials with very high surface areas, as a
while still attached [3–6]. Abnormal action on insoluble consequence of very small pore sizes, or even with a mobile
substrates may also be relevant in certain disease states gel-type network without well-defined pores. Thus there is
[7–9], such as attack on cartilage in arthritis and cell surfaces a clear possibility that enzyme molecules will not be able
in pancreatitis. to penetrate all parts of the support bead, or at least their
When enzymes are employed in solid-phase synthesis, it diffusion through the particle will be greatly hindered. One
is normally observed that rates and yields are substantially type of particle that has been particularly recommended for
reduced when the usual substrates are covalently attached to enzyme reactions are the PEGA [poly(ethylene glycol) acryl-
a solid particle. Enzyme assays on solid-phase substrates are amide] range . These consist of a core acrylamide structure
usually used just to give relative rates, but where absolute ac- to which poly(ethylene glycol) chains are attached. The beads
tivity has been measured, it is again low compared with those swell in both water and a range of organic solvents, and permit
in solution. There have recently been some initial approaches high loadings of reactive sites at approx. 200 µmol · g−1 .
to improve understanding of how enzymes behave in such There is evidence that PEGA beads may restrict penetration
a system . Better understanding should be the basis for of enzyme molecules, particularly larger ones. One of the
designing reaction systems to obtain better rates and yields. clearest demonstrations used confocal Raman microscopy to
It will also aid the interpretation of results obtained in such monitor reaction in the bead interior .
An alternative method to image the sites of enzyme reac-
Key words: confocal microscopy, ﬂuorophore, poly(ethylene glycol) acrylamide (PEGA), solid- tion is to use fluorescence microscopy. Ideally, we would like
phase synthesis, surface chemistry.
to use optical sectioning of intact beads, to avoid possible arti-
Abbreviations used: Ac, acetyl; CPG, controlled pore glass; PEGA, poly(ethylene glycol)
acrylamide. facts from sample preparation. Confocal fluorescence micro-
email P.J.Halling@strath.ac.uk scopy is not normally applicable to these systems, however,
C 2006 Biochemical Society
310 Biochemical Society Transactions (2006) Volume 34, part 2
Figure 1 Spatially resolved time course of thermolysin action on Scheme 1 Equilibrium positions for coupling of saturated
immobilized substrate in PEGA beads aqueous Fmoc (ﬂuoren-9-ylmethoxycarbony)-Gly to free and
Fmoc (ﬂuoren-9-ylmethoxycarbonyl)-Phe-Phe-PEGA beads were ex- immobilized phenylalanine residues
posed to enzyme for the times shown, then amino groups released were Reaction at bulk pH 7.4 was catalysed by thermolysin, with PEGA1900 as
labelled with dansyl. Optical sections of intact beads were examined by solid phase. Data from .
two-photon ﬂuorescence microscopy. Reproduced from . c 2003
The Royal Society of Chemistry.
reaction at different positions in the optical section through
the bead. This has proved quite challenging to implement as a
quantitative assay, but we are now close to success ( J. Deere,
A. Lalaouni, B. Maltman, G. McConnell, J. Girkin, S.L.
Flitsch and P.J. Halling, unpublished work).
Equilibrium of reactions on solid-attached
as the high fluorophore concentrations absorb excitation substrates
light before it can reach deep into the bead. As a result, the On first thought, it might be expected that attachment to
fluorescence detected from the bead core is ‘artifactually’ low. a solid support would have little effect on the equilibrium
Instead, we have used two-photon fluorescence microscopy of a reaction, provided the medium stays the same (e.g.
for this purpose. In this approach, light absorption is largely aqueous). In fact, however, there can be a substantial shift in
restricted to the focal point as this is scanned through the equilibrium position, for example making peptide synthesis
bead, because only there is the intensity high enough for largely favourable even in an aqueous environment (e.g.
substantial two-photon absorption events. We were able to Scheme 1; [15,16]). This can be exploited preparatively,
show that two-photon microscopy gave fairly good quantifi- but may also be the reason for poor yields observed in
cation of fluorophore concentrations throughout the bead, some attempted hydrolysis reactions on peptide and other
after correction for an aberration effect as a function of depth substrates.
into the sample . Three factors may contribute to the shift in equilibrium
A first approach to identify where the enzyme was acting .
used a post-reaction labelling method . Enzyme-catalysed (i) With one component of a coupling reaction attached to
hydrolysis of a peptide bond liberated a free amino group the solid phase, it is easier to add and remove a large excess
on the substrate portion still bound to the bead. After of the other. This is of course theoretically trivial, although a
stopping the enzyme reaction at various times, the free amino practically useful feature of solid-state chemistry. However,
groups were labelled by reaction with dansyl chloride. After it is not the explanation for some observed shifts, such as
washing, imaging of the positions of the dansyl fluoro- the example above, because this mass action effect has been
phore revealed where the enzymatic reaction had occurred allowed for.
(Figure 1). This indicated that the reaction in the bead core (ii) If the soluble reaction component contains hydro-
lagged significantly behind that in the surface regions. In turn, phobic groups (e.g. the side chain or protecting group of an
this observation suggests that the effective diffusion coeffi- amino acid), this can contribute a large effect. The synthesis
cient of the enzyme molecules inside the bead is considerably reaction leads to transfer of these groups out of solution
lower compared with that in aqueous solution. and into an environment near the solid surface, where
It is preferable to be able to monitor the reaction con- hydrophobic hydration is less unfavourable. After reaction,
tinuously in real time. For this purpose, we have used an these groups will be less in contact with water, and this
alternative strategy (which has also been adopted to probe water will already be partly ordered by the nearby surface,
reaction at different spots on a plane surface [5,6]). An amino- or by polymer chains that form part of the support structure.
coumarin fluorophore is first attached to the solid support This effect is obviously greatest for the most hydrophobic
through a carboxylic acid side chain that does not interfere reactants, and for benzyloxycarbonyl (Z) phenylalanine can
with fluorescence. Then, the amino group is acylated with an be as large as 104 (change in ratio of amide to amine).
enzyme-cleavable residue such as Ac (acetyl)-Phe. The amide (iii) Mutual electrostatic repulsion of charged groups on
produced is non-fluorescent, so the bead now becomes rel- the solid surface will favour synthesis. In amide hydrolysis,
atively dark. Enzymatic removal of the Ac-Phe group, for ex- for example, ionization of the liberated amino and carboxy
ample by chymotrypsin, liberates the free amino group again, groups makes a major contribution in favouring the reaction
and restores fluorescence. Hence monitoring the develop- in water. If the ionizable groups are attached to the solid
ment of fluorescence gives a real-time probe of the rate of particle, their ionization will be suppressed by mutual
C 2006 Biochemical Society
repulsion, and hence synthesis will be relatively favoured over Figure 2 Effect of oligoglycine spacer length on rate of
hydrolysis. The effect can be seen as a change in pK of the chymotrypsin-catalysed hydrolysis
surface-attached groups (in terms of the bulk pH), but is From none to eight glycine residues were attached to PEGA or amino-
properly modelled as a Donnan equilibrium. The effect propylsilane-treated CPG, followed by N-terminal Ac-Trp. The measured
is largest at low ionic strength, and is relatively small in the rate of Ac-Trp release after addition of chymotrypsin to medium (66 µg ·
case studied with approx. 0.1 M salt. ml−1 ) is shown. Rates are expressed as percentage of maximum, in
nmol · (g of support)−1 · min−1 , of 113 (CPG) and 1.56 (PEGA). (J. Deere,
Effect of spacer length attaching substrate A. Lalaouni, L.F. Solares, S.L. Flitsch and P.J. Halling, unpublished work).
moiety to surface
Enzyme catalysis requires binding of the substrate moiety in
the active site. Obviously, the presence of the support surface
may interfere in the process, particularly if this is only a short
distance from the site of reaction in the substrate. With this
concern in mind, some workers have attached the substrate
moiety using a ‘spacer’, a relatively flexible chain that takes
the site of reaction further away from the surface . The
use of such spacers draws upon successful approaches used
in affinity chromatography and immobilized enzymes. We
have recently made a systematic study of the effects of spacer
length on chymotrypsin action (J. Deere, A. Lalaouni, L.
Solares, S.L. Flitsch and P.J. Halling, unpublished work). Two
support materials were used, PEGA and CPG (controlled glycine residues, which may be related to the known change
pore glass). Amino groups are present in the former at the in conformation of oligoglycines in solution.
end of the poly(ethylene glycol) chains, and were introduced
on the latter using an aminopropyl silane coupling agent.
Coupling of oligoglycine chains on to these amino groups References
1 Kadereit, D. and Waldmann, H. (2001) Chem. Rev. 101, 3367–3396
formed the spacers, with final attachment of an Ac-Trp
2 Reents, R., Jeyaraj, D.A. and Waldmann, H. (2001) Adv. Synth. Catal.
residue. Chymotrypsin in solution catalysed the release of 343, 501–513
free Ac-Trp, and the rate of its appearance in the supernatant 3 Lee, Y.S. and Mrksich, M. (2002) Trends Biotechnol. 20, S14–S18
4 Min, D.H. and Mrksich, M. (2004) Curr. Opin. Chem. Biol. 8, 554–558
was determined by HPLC analysis. The relatively small
5 Salisbury, C.M., Maly, D.J. and Ellman, J.A. (2002) J. Am. Chem. Soc. 124,
concentrations of Ac-Trp make rate estimations less accurate 14868–14870
than in solution, but some features of the pattern are clear 6 Zhu, Q., Uttamchandani, M., Li, D.B., Lesaicherre, M.L. and Yao, S.Q.
(2003) Org. Lett. 5, 1257–1260
(Figure 2). In the absence of any spacer, the rates are
7 Berdowska, I. (2004) Clin. Chim. Acta 342, 41–69
relatively low, particularly on the PEGA resin. This probably 8 Dickinson, D.P. (2002) Crit. Rev. Oral Biol. Med. 13, 238–275
reflects the nature of the amine leaving group as well as 9 Yasuda, Y., Kaleta, J. and Bromme, D. (2005) Adv. Drug Deliv. Rev. 57,
interference from the support surface. On PEGA, there is
10 Halling, P.J., Ulijn, R.V. and Flitsch, S.L. (2005) Curr. Opin. Biotechnol. 16,
a somewhat hindered 2-aminopropyl group, while the 1- 385–392
aminopropyl group on CPG may be disfavoured relative 11 Meldal, M. (2002) Biopolymers 66, 93–100
12 Kress, J., Zanaletti, R., Amour, A., Ladlow, M., Frey, J.G. and Bradley, M.
to an amino acid residue. When an oligoglycine chain is
(2002) Chem. Eur. J. 8, 3769–3772
introduced, two residues are sufficient for maximal rates on 13 Ulijn, R.V., Brazendale, I., Margetts, G., Flitsch, S.L., McConnell, G.,
PEGA, while at least four are required for the optimum with Girkin, J. and Halling, P.J. (2003) J. Comb. Chem. 5, 215–217
14 Bosma, A.Y., Ulijn, R.V., McConnell, G., Girkin, J., Halling, P.J. and Flitsch,
CPG. This probably reflects the flexibility offered by the
S.L. (2003) Chem. Commun., 2790–2791
long poly(ethylene glycol) chains in PEGA, to the end of ˜
15 Ulijn, R.V., Baragana, B., Halling, P.J. and Flitsch, S.L. (2002) J. Am.
which the oligoglycine is attached. In contrast with CPG, Chem. Soc. 124, 10988–10989
16 Ulijn, R.V., Bisek, N., Halling, P.J. and Flitsch, S.L. (2003)
the amino group of the silane is close to and probably fairly
Org. Biomol. Chem. 1, 1277–1281
fixed relative to the rigid glass surface. Hence more glycine 17 Schmitz, C. and Reetz, M.T. (1999) Org. Lett. 1, 1729–1731
residues are needed for optimal flexibility at the reaction site.
In both cases, there is a decline in rate with more than six Received 21 November 2005
C 2006 Biochemical Society