Technical Limitations

Using state-of-the-art instrumentation, force resolution in SMFS is limited only by the thermal fluctuations detected by the force sensor, ttus, the cantilever is a critical element in SMFS because its mechanical properties dictate the force sensitivity, response times, and spatial resolution. According to the so-called fluctuation-dissipation theorem, a reduction in the size of the force sensor should improve the signal-to-noise ratio (Gittes and Schmidt 1998). tte signal-to-noise ratio of force measurements is independent of the stiffness of the cantilever but can be improved by reducing its dimensions, as has been verified experimentally. As a result, smaller but still soft cantilevers allow for higher sensitivity, faster response times, and a high spatial resolution (Viani et al. 1999). ttis new generation of cantilevers is a promising development permitting resolutions of 7-10 pN (e.g., Schwaiger et al. 2002). However, they are currently expensive, as they have to be custom-made using micromachining techniques, ttey also require a modified SMFS head, in order to avoid optical interference from some laser light spilling over the cantilever and reflecting off the substrate. At present, although SMFS experiments can readily resolve single amino acid length changes (Carrion-Vazquez et al. 1999b), they cannot provide detailed structural information yet. With the further development of smaller cantilevers, the secondary structure of proteins and single hydrogen bonds may be resolved in the future, although some theoretical considerations have been raised against this possibility (ttompson and Siggia 1995). If we extrapolate to the size limit, the cantilever should be replaced by a single elastic molecule to increase the precision of this method. A first step in this direction has been the development of an all-or-none force sensor based on a short DNA duplex (Albrecht et al. 2003). In the future we may also see the development of molecular force transducers with much better force resolution and capable of being used for in vivo applications (see later).

Table 8.2. Mechanical properties of a selection of protein-biomolecule pairs studied by SMFS

Interaction pair

Fu (pN)

vr (nm ms"1)

Xp(nm)

Kfd s-1)

Thermodynamic data

Attachment (T: tip; S: surface)

Single unbinding event criterior

Controls and experiment design

Remarkable aspects of the work

References

Avidin-biotin

160±20

0.93a

AH=21.5 kcal moM; &G= -20.4 kcal mor1

T: receptor via biotin-labeled BSA; S: agarose beads with ligands

Quanta 1 behavior

Blocking with avidin, elastic agarose beads, different ligands/receptor, thermodynamical data, changes in pH

Pioneer work on interaction studies

Florin etal. (1994); Moy et al. (1994)b

HAS-antiHSA*

240±48

0.2 (r=54 nN s_1)

6

6.7 x10-4

Ka=10nM

T: Ab via PEG; S: Ag attached to mica via PEG

Spacer linkers

Spacer linker, blocking with free HSA, PEG spacer to control unspecific interactions

Spacer linkers, /c0ff directly measured

Hinterdorfer et al.

BSA-IgG

antibiotin*

60 ±20

Tand Sto Ag and Ab via DSU

Quanta 1 behavior

Blocking with free streptavidin and biotin, changes in the pH, BSA instead of Ag, (-) control antibody,switching tip/ surface

One of the first Ag-Ab AFM studies

Dammer et al. (1996)

Ferritin-

antiferritin lgG2a*

49±10

-

-

-

-

T: covalently to ferritin; S: Ab via glutaraldehyde

Quanta 1 behavior

Control antibody, polyepitopic antigen, blocking with ferritin

Using a polyepitopic antigen

Allen etal.(1997)d

Fluorescein-Ab scFv fragments*

50±4

1

AG=-12.14 kcal mol-1; Kd=0.75 nM

T: fluorescein via PEG; S: Ab bound to gold by Cys

Spacer linker, AFM imaging

Spacer linker, AFM imaging, blocking fluorescein, mutant scFv, thermodynamics data

AFM imaging, mutants

Ros eta 1.(1998)e

Proteoglycan of Microciona proliféra+

40±15

Covalently bound to TandS

Quanta 1 behavior

Changes in Ca2+, Mg2+ as control, blocking with Ab, AFM and EM imaging

Pioneer AFM study of interaction forces

Dammer et al. (1995)

GRGDSP peptide1

42 ±4

1

T: peptides via PEG; S: 0^3 expressing osteoclasts

Quanta 1 behavior

Spacer linker, control peptides,control cells, changes in pH,blocking with peptides

First study using intact cells

Lehenkari and Horton (1999)

Interaction pair

MpN)

vr (nm

Xp (nm)

koffi s-1)

Thermo

Attachment (T:

Single

Controls and experiment

Remarkable

References

ms"1)

dynamic data

tip; S: surface)

unbinding

design

aspects of the

event

work

criterion

VE-cadherin

40

0.8

0.59

1.8

Kd=

T: VE-cadherin

Unimodal

Spacer linker, no calcium,

Lateral

Baumgartner et al.

(homophilic

1.15 nM

via PEG; S:

Gaussian fit

blocking with Ab, changes

interaction,

(2000)

interaction)*

VE-cadherin

in speed

covalently via

multiple

PEG

speeds

LFA-1 -

20-320

0.1-15

0.018-

4-57

_

T:T cells via

Unimodal

Blocking with antibodies,

Multiple r:

Zhang et a I. (2002)f

ICAM-1+

(r=20-

0.21

Con A; S:

Gaussian fit

changes in Mg2+; LFA-1

two loading

50,000

soluble ICAM-1

varieties

regimes

pN s"1)

adsorbed to

mica surface

P-selectin-

5-50

0.25

0.46

0.22

_

T: PSGL-1 via

Comparison

Monomeric/dimeric PSGL-

Catch bonds

Marshall et a I. (2003)

PSGL-1+

Abs; S: lipid

between

1, Ab as ligand, blocking

vs. slip bonds

bi layers with

monomeric

with Ab, changes in free

P-selectin

and dimeric

Ca2+, flow chamber

data

GroEL-

48±20

0.12

2.5

-

-

T: pepsin via

"Compre

AFM imaging, blocking

Compression-

Sekiguchietal. (2002,

denatured

(ATP);

±3.4

Sulfo-LC-SPDP;

ssion-free"

with pepsin, compression-

free method

2003)

pepsin*

44±20

(ATP);

method,

free method, changes

(no ATP)

9.2±

S: GroEL

unimodal

in ATP

5.9

adsorbed to

Gaussian fit

(no ATP)

mica surface

Bacterio-

100-200

0.04

-

Nonspecific

AFM

AFM imaging, comparison

Integral

Müller et a I. (2002)h

rhodopsin-

attachment (S:

imaging,

of expected length with

membrane

native purple

native purple

comparison

length of helices, changes

protein,

membrane5

membrane

with

in pH, retinal configuration

combination

3.2g

adsorbed to

unfolded

changes

of AFM

Helix-pair ED

167±20

1.0x10-2g

mica)

protein

imaging

Helix-pair BC

99±16

8.69

3.4x10_5g

length

and force

spectroscopy

Interaction pair Fu (pN) vr (nm Xß(nm) /f0ff(s ) Thermo- Attachment (T: Single Controls and experiment Remarkable References ms"1) dynamic data tip; S: surface) unbinding design aspects of the event work criterion

ExpG-target DNAs'p

50 -165

T: DNA via Unimodal Functionalized AFM tip First

PEG;S:ExpG Gaussian fit and clean surface, blocking protein-DNA

attached to with free DNA, comparison interaction silanized mica with EMSA AFM study,

Bartels et al. (2003)

Lex A-target DNAs*

LexA-recA

Variable (depending on r)

T: DNA via CMA; Reducing Spacer linker, fluorescence Changes in Kühner et al. (2004)

S:TFviaCMA success rate assays, CMA to control r, changes in unspecific interactions, spacer length changes in spacer

LexA-yebG

interaction pairs are classified as avidin-biotin, antigen (Zlg)-antibody (Ab) (*), adhesion proteins (t), cytoplasmatic proteins (#), protein-lipids (§), and protein-DNA pP). Calculated as-AH/Fu bRelated articles, Izrailevet al. (1997) (steered molecular dynamics simulation of biotin-avidin), Wong et al. (1998), Merkel et al. (1999) "Related article, Wielert-Badt et al. (2002)

dRelated article, Allen et al. (1999) (discussion on the influence of epitope availability in AFM studies ofthe Ag-Ab interaction).

eRelated article, Schwesinger et al. (2000) (the correlation between unfolding force and k0ff is shown here studying antibody mutants and using different loading rates).

fThis methodology has been also applied to study the interactions between fibronectin and a5ß! integrin (Li et al. 2003) and P-selectin and its ligand (Hanley et al. 2003). 9Data from Janovjaket al. (2004)

hRelated articles, Oesterhelt et al. (2000) (first published report on the subject), Janovjak et al. (2003) (temperature dependence ofthe process), Janovjaket al. (2004) (energy landscape)

tte future use of SMFS in proteomics will require new developments to enhance the speed of analysis and sample throughput, so that parallel analysis of samples can be efficiently performed. In addition to the development of small cantilevers, there are other steps in this direction: the "micro SMFS" (based on a multibridged cantilever; Chu et al. 2000), the so-called tip arrays (Minne et al. 1998), and the "Millipede" device (Vettiger et al. 2002). Besides increasing the throughput, multiple cantilevers have the advantage over single cantilevers that reference cantilevers can be used to effectively remove thermal drifts, mechanical disturbances coming from noise (such as vibrations), and false positives.

Current technology does not permit SMFS to directly measure the forces involved in unfolding/folding inside cells, although it has already been used to measure the mechanical properties of the extracellular portions of interacting proteins located on the plasma membrane of living cells (Benoit and Gaub 2002). tte ultimate goal of protein mechanics is to fully access the machinery of living cells. Several avenues are being explored to achieve this, including the use of magnetic beads or fluorophores. tte recent development of a hybrid AFM/total internal reflection fluorescence microscopy apparatus (Sarkar et al. 2004) and the characterization of the mechanical properties of GFP (Dietz and Rief 2004) are the first steps towards the development of a molecular force sensor based on fluorescence that could be fused to the protein of interest by genetic engineering. Many alternative approaches are also being explored at present, as the biosensor field is a very active area of research nowadays.

Two additional constraints limit SMFS potential: the lack of knowledge of the initial state of the adsorbed sample and the fact that there is no real choice of the specific system (molecules or other entities to be analyzed). As we have mentioned, SMFS is typically a "blind" fishing expedition where single-molecule pickup is random. We cannot choose generally what we get in each event and the options are many since the immobilized/adsorbed protein sample may consist of a heterogeneous mixture of species including native, partially denatured, completely denatured, and misfolded single molecules, as well as aggregates and native supramolecular complexes, tterefore, AFM imaging of the sample prior to SMFS would be of paramount importance to "have a look" before and after performing the pulling experiment. In this way, specific molecules or other entities could be chosen and pulled at specific points. Moreover, this would also enable specific immobilization systems to be developed, to better preserve protein conformation (as mentioned, physisorp-tion can often cause partial or total denaturation of the protein). However, with the exception of a few amenable systems such as bacteriorhodopsin, which forms rigid 2D crystals in the purple membrane (Sect. 8.5.2), AFM imaging/SMFS analyses in parallel (this method is also known as "simultaneous," sensu lato) are rarely seen (Meadows et al. 2003). tte main reason for this is that proteins are soft materials that are "hard to image" using classical imaging modes (i.e., contact and "tapping"). In this sense, the development of the so-called jumping mode, as it is milder on the sample, has generated some expectation for this particular application (de Pablo et al. 1998). Furthermore, a compromise in the experimental conditions must be established, since the requirements of each of the two AFM configurations are typically different: relatively thick layers and mild attachment to substrate (typically gold) in

SMFS vs. monolayers and strong attachment to substrate (usually mica) in AFM imaging.

It was the advent of polyprotein production that moved the field of intramolecular protein mechanics forwards by providing unequivocal single-molecule fingerprinting. Development of a similar tool (of universal applicability to proteins) to unmistakably identify single unbinding events would be desirable to provide a further impulse to the field of intermolecular protein mechanics. Still, since the methods available for constructing polyproteins are very laborious, a means to simplify this process would also be welcomed, ttus, the lengthy construction of the concatemer, the main bottleneck in the most widely used of those methods (i.e., the recombinant polyprotein), should be shortened. Implementing a single-step method to translate the concatemer in vitro would also be useful (expression in vivo, in Escherichia coli, remains the norm) as only small amounts are typically needed to complete the mechanical analysis of a polyprotein (a few tens of micrograms of protein), ttis method would give us more control of the expression conditions, which may help to overcome possible problems related to toxicity and solubility of the protein. Recent improvements of in vitro expression systems, both in terms of yield and in the length of the protein that can be produced, offer some promise for the future simplification of this step.

Was this article helpful?

0 0

Post a comment