As we have seen (Sect. 8.3.2), proteins tend to denature to some extent when adsorbed onto a surface, and for this reason long modular proteins are very good systems for SMFS studies. Such proteins are made up of several modules and, thus, two of these can readily be partially or totally "sacrificed" for the required attachments of the molecule to the cantilever tip and to the substrate. Furthermore, the "proximal region" to the substrate of a force-extension curve is often the main source of nonspecific interactions and spans typically the first 30-70 nm (Fig. 8.4b). tterefore, the forces involved in extending short proteins or a typical module (even if they were not denatured by the adsorption required for the attachments) may be masked by the "noise" caused by such nonspecific interactions, ttus, if we also consider than modular proteins tend to have a mechanical function, it is not surprising that the earliest SMFS experiments (and most of the SMFS studies published to date) focused on multidomain mechanical proteins (native or recombinant) such as titin, tenascin, spectrin, and fibronectin (Oberhauser et al. 1998, 2000, 2002; Rief et al. 1997b, 1999). However, modular proteins are composed of heterogeneous populations of modules (with different sizes, sequences, structures, and stabilities); therefore it is not possible to identify which force peaks correspond to which modules in the resulting sawtooth pattern.
ttus, although these studies provide a simple way of first approaching the study of the global mechanical properties of a modular protein, it seems clear that pure-periodic proteins (i.e., perfectly repetitive), the so-called polyproteins, would be an ideal choice system to allow the study of the mechanical properties of a single protein domain (or even a nonmodular protein). However, there are only a few naturally occurring polyproteins that are composed of identical repeats (at the amino acid level) without having any spacers between them, such as polyubiquitins and the yeast a-factor. Indeed, in most cases the repeats are not perfect or are linked by spacer sequences of varying lengths (Finley and Varshavsky 1985; Wiborg et al. 1985 and references therein), tte advent of methods to synthesize homomeric polyproteins (i.e., head-to-tail covalently linked tandem repeats of a single domain or protein; referred to in this review as polyproteins, sensu lato) opened the way to the mechanical analysis of single modules and nonmodular proteins (Carrion-Vazquez et al. 1999a, 2000). Since then, several additional strategies have been established for the artificial synthesis of polyproteins for SMFS experiments. All the strategies developed so far fall into two basic categories:
1. Methods based on the construction of repeats at the DNA level through genetic engineering (Carrion-Vazquez et al. 1999a). In order to permit directional head-to-tail cloning, two strategies were originally used: one using a single nonpalin-dromic (asymmetrical) restriction endonuclease recognition site (AvaI), which adds a Leu-Gly linker; the other uses palindromic (symmetrical) restriction sites and two restriction enzymes with compatible cohesive ends (BamHl/Bglll and Kpnl), which add an Arg-Ser linker. Sequencing is typically carried out at the monomer level using this method, before concatemerization, although the cloned concatemer could in principle be sequenced either directly by deletion sequencing or after its digestion to monomers (and processing an "oversampled" number of clones). A "cassette" strategy has recently been established based on the use of multiple restriction sites. Although this method generates a polyprotein with different amino acid linkers, the sequencing of the concatemer is easier and it permits the selective introduction of proteins or domains at specific positions (Steward et al. 2002). 2. A method based on the synthesis of repeats directly at the protein level, by disul-fide-bonding cysteine residues (Yang et al. 2000). Using this method, directionality (i.e., head-to-tail links) in the assembling of monomers is only possible if the synthesis is performed in the solid state.
Polyproteins provide multiple advantages for SMFS studies (Carrion-Vazquez et al. 2000; Fig. 8.5). tteir monotonous nature for the first time permitted singlemodule analysis and the experimental demonstration of the stochastic nature of the process of mechanical unfolding, ttese systems are also an ideal choice for force spectroscopy experiments because their periodical design provides a way to sort out nonspecific interactions and to unequivocally identify single-molecule events (i.e., single-molecule "fingerprinting"). Furthermore, some repetitions of the polyprotein can have "sacrificial" roles: they can be used for protein attachment (through partial or complete denaturation), they can act as polymer spacers that physically separate most of the repeats from the proximal region of the substrate, and they can also take up the pressure denaturation effects that may likely result from the high forces applied to attach the polyprotein to the cantilever tip (while in the nanonewton range, these pressures are estimated to range up to several gigapascals). tte use of polyproteins also permits some spatial features to be amplified, which has enabled the detection of small amino acid insertions, conformational intermediates that are spatially very close to the native state, as well as aberrant folding events. In the case of unstructured proteins, which present featureless force-extension curves (e.g., PEVK or N2B titin domains, see later), the construction of a chimeric polyprotein, which includes a marker domain (e.g., 127 titin domain; see later) fused to the protein under study, has enabled single-molecule events to be unequivocally fingerprinted (Li et al. 2001, 2002). In the case of "structured proteins" the use of heteromeric polyproteins (instead of homomeric polyproteins) allows the inclusion of a convenient internal standard (e.g., 127 module), covalently built in the molecule, which may also permit the "solubilization" of proteins that are insoluble as homomeric polyproteins (as shown for the 127 module; Forman et al. 2005 and references therein). Lastly, because they are perfect repetitions of the system under study, polyproteins have the added bonus of significantly increasing the rate of data acquisition in SMFS experiments. An alternative strategy to the use of homomeric and heteromeric polyproteins in SMFS is "to sandwich" the protein under study between two reference
Fig. 8.5. The use of polyproteins in SMFS. a Mechanical analysis of single protein modules. The force-extension curve of a modular protein (recombinant 127-134 Ig region of I band of titin) typically shows a steady rise in the force peaks, indicating the hierarchical unfolding of its modules. However, none of the peaks in the spectrum can be specifically attributed to any module. On the other hand, stretching an 127 polyprotein produces a sawtooth pattern in which the force fluctuates around an average value. This permits an individual module to be characterized (modified from Li et al. 2000b). All the sawtooth pattern recordings in the
Fig. 8.5. The use of polyproteins in SMFS. a Mechanical analysis of single protein modules. The force-extension curve of a modular protein (recombinant 127-134 Ig region of I band of titin) typically shows a steady rise in the force peaks, indicating the hierarchical unfolding of its modules. However, none of the peaks in the spectrum can be specifically attributed to any module. On the other hand, stretching an 127 polyprotein produces a sawtooth pattern in which the force fluctuates around an average value. This permits an individual module to be characterized (modified from Li et al. 2000b). All the sawtooth pattern recordings in the polyproteins made out of either the same module or two different modules (see Table 8.1 for examples and references).
Nevertheless, in spite of all the advantages of using polyproteins in SMFS studies, the technique in its current form is based on a trial-and-error approach in which both the attachment points of the polyprotein molecules and the number of modules trapped in each trial are random.
remaining panels ofthe figure were obtained using 127 polyproteins. b Demonstration of the stochastic nature of mechanical unfolding. Forced unfolding experiments on polyproteins generate a distribution of forces rather than a specific force value, reflecting the dominant role of thermal forces at the single-molecule level (modified from Carrion-Vazquez et al. 1999b). c Polyproteins also act as spacer arms, increasing the length ofthe protein beyond the region of the nonspecific interactions between the cantilever tip and the polyprotein layer (which typically spans about 30-70 nm).The pseudoperiodicity ofthe sawtooth pattern identifies single domain unfolding (modified from Carrion-Vazquez etal. 2000). d Amplification of mechanical properties. Top: when compared with the wild type (black trace), a five glycine residue insertion in a mutant polyprotein (red trace) elongates the contour length ofthe molecule with each unfolding event (1.91 nm per module) such that it can be readily detected from the third or fourth force peak ofthe sawtooth pattern (after Carrion-Vazquez et al. 1999b). Bottom: an unfolding intermediate originating from the rupture of a patch of hydrogen bonds between the A and B p-strands of 127 (Fig. 8.9) is detected as a deviation ("hump") ofthe WLC (blue lines). In the polyprotein analyzed in each experiment, the first peak reflects the rupture of the AB patch of all the modules trapped in this stretch (after Fisher et al. 2000). e Detection of misfolding events. During repeated unfolding-refolding cycles, some sawtooth patterns (less than 2%) displayed missing unfolding force peaks ("skips"), which were interpreted as originating from the rupture of "superfolds" formed during the refolding process. This interpretation was based on the observation that the increment in the contour length was longer than the sum ofthe lengths of two consecutive peaks in the unfolding pattern, and the difference accounted for the stretching ofthe linkers that separate two consecutive modules (from Fisher et al. 2000;0berhauser et al. 1999; with permission, copyright 1999 Macmillan Publishers Ltd.). f Single-molecule fingerprinting with polyproteins. The mechanical properties of "unstructured" (i.e., without tertiary structure) proteins or domains can be done using polyproteins as an internal marker for single-molecule identification. 1 The "sandwich" method: Force-extension curve of a protein chimera containing the N2B region from human cardiac titin flanked on either side by three 127 domains (I273-N2B-I273). 2 "Alternating copolymer" method: force-extension curve of a protein chimera containing three PEVKdomains from human cardiac titin, alternated with three Ig 127 domains (I27-PEVK)3. In both curves, a nonlinear least-squares fit (i.e., a Levenberg-Marquardt curve fit) ofthe WLC equation (thin line) to the force-extension curve prior to the first 127 unfolding event was used to measure the contour length, La and persistence length, p, both ofPEVKas well as ofN2B (after Li et al. 2002)
Was this article helpful?