Analytica Chimica Acta, accepted, 07/07/2015. This is the accepted version without proofing
corrections. DOI: 10.1016/j.aca.2015.06.011
.
Page 2 of 26
1. Introduction.
Rapid, accurate characterization of macromolecule structure, stability, and
aggregation is a critical/challenging task in analytical science. To provide real time
information regarding protein stability, the analytical method should be rapid, require a
minimum of sample handling and provide high-information content. Spectroscopic methods
can fulfil most of these requirements and fluorescence spectroscopy is widely used [1-5].
The use of intrinsic fluorescence is attractive because it facilitates analysis of proteins in their
native state without the use of extrinsic labels, which involves additional sample
manipulation and may also change macromolecule structural/chemical properties. For
example, the hydrodynamic radius of HSA at 25 °C was variously reported to be between
31.5 and 35.6 Å [6-8] in the native state, and 39.1±2.5 Å with a TMR label [9].
One of the most widely studied and important proteins involved in human health is
HSA because of its abundance and role as a transport protein [10-12]. HSA is comprised of
585 amino acid residues of which phenylalanine (Phe), tryptophan (Trp), and tyrosine (Tyr)
are fluorescent [13, 14]. Structurally HSA is composed of three similar α-helical domains IIII, each with a pair of subdomains A and B (See Supplementary information (SI), Fig. S-1)
[15]. A single Trp, Trp-214, located in subdomain IIA dominates emission [16]. Intrinsic
fluorescence has been widely used to monitor structural changes in HSA (steady-state or
time-resolved emission, or anisotropy) [16-19]. These studies have shown that when
thermally denatured, intermediate unfolded states were observed, and that each domain
unfolded independently [9, 17, 20]. After thermal denaturation at 50–60 °C, refolding on
cooling lead to less stable, native like structures. Above 60°C, unfolding became irreversible,
so that when cooled, HSA refolded into structures different from the native state [17, 18, 20,
21].
Multidimensional fluorescence spectroscopy (MDF) techniques, like excitationemission matrix (EEM) [22] and TSFS [23], have been successfully used for simultaneous
analysis of multi-fluorophore mixtures for a wide range of applications [24-28]. In MDF
spectroscopy, the data map generated is in essence a spectral fingerprint of the multiple
fluorophores present in a protein or a complex sample [26, 28]. HSA fluorescence comprises
of overlapping emission from 31 Phe, 1 Trp, and 18 Tyr residues. Fluorophore spectral
properties are influenced by a variety of factors linked to protein structure, e.g. Förster
resonance energy transfers (FRET), electron transfer quenching by peptide bonds, variable
solvent exposure, and inner filter effects (IFE) [29, 30]. More specifically the precise
fluorophore location and separation between fluorophores and quenchers are the critical
parameters. The combination of all these factors determines the gross topography of the
emission space and thus generates a unique MDF spectrum. The use of EEM for observing
HSA structural changes during freezing has been demonstrated [31, 32], but multivariate
analysis was not used to better understand the observed spectral changes in terms of the
individual fluorophores.
One of the main problems with fluorescence spectroscopy in general is that the broad
excitation and emission bands result in extensive spectral overlap between the chromophores
and fluorophores in complex mixtures. Chemometric techniques can be used to resolve
corrections. DOI: 10.1016/j.aca.2015.06.011
.
Page 2 of 26
1. Introduction.
Rapid, accurate characterization of macromolecule structure, stability, and
aggregation is a critical/challenging task in analytical science. To provide real time
information regarding protein stability, the analytical method should be rapid, require a
minimum of sample handling and provide high-information content. Spectroscopic methods
can fulfil most of these requirements and fluorescence spectroscopy is widely used [1-5].
The use of intrinsic fluorescence is attractive because it facilitates analysis of proteins in their
native state without the use of extrinsic labels, which involves additional sample
manipulation and may also change macromolecule structural/chemical properties. For
example, the hydrodynamic radius of HSA at 25 °C was variously reported to be between
31.5 and 35.6 Å [6-8] in the native state, and 39.1±2.5 Å with a TMR label [9].
One of the most widely studied and important proteins involved in human health is
HSA because of its abundance and role as a transport protein [10-12]. HSA is comprised of
585 amino acid residues of which phenylalanine (Phe), tryptophan (Trp), and tyrosine (Tyr)
are fluorescent [13, 14]. Structurally HSA is composed of three similar α-helical domains IIII, each with a pair of subdomains A and B (See Supplementary information (SI), Fig. S-1)
[15]. A single Trp, Trp-214, located in subdomain IIA dominates emission [16]. Intrinsic
fluorescence has been widely used to monitor structural changes in HSA (steady-state or
time-resolved emission, or anisotropy) [16-19]. These studies have shown that when
thermally denatured, intermediate unfolded states were observed, and that each domain
unfolded independently [9, 17, 20]. After thermal denaturation at 50–60 °C, refolding on
cooling lead to less stable, native like structures. Above 60°C, unfolding became irreversible,
so that when cooled, HSA refolded into structures different from the native state [17, 18, 20,
21].
Multidimensional fluorescence spectroscopy (MDF) techniques, like excitationemission matrix (EEM) [22] and TSFS [23], have been successfully used for simultaneous
analysis of multi-fluorophore mixtures for a wide range of applications [24-28]. In MDF
spectroscopy, the data map generated is in essence a spectral fingerprint of the multiple
fluorophores present in a protein or a complex sample [26, 28]. HSA fluorescence comprises
of overlapping emission from 31 Phe, 1 Trp, and 18 Tyr residues. Fluorophore spectral
properties are influenced by a variety of factors linked to protein structure, e.g. Förster
resonance energy transfers (FRET), electron transfer quenching by peptide bonds, variable
solvent exposure, and inner filter effects (IFE) [29, 30]. More specifically the precise
fluorophore location and separation between fluorophores and quenchers are the critical
parameters. The combination of all these factors determines the gross topography of the
emission space and thus generates a unique MDF spectrum. The use of EEM for observing
HSA structural changes during freezing has been demonstrated [31, 32], but multivariate
analysis was not used to better understand the observed spectral changes in terms of the
individual fluorophores.
One of the main problems with fluorescence spectroscopy in general is that the broad
excitation and emission bands result in extensive spectral overlap between the chromophores
and fluorophores in complex mixtures. Chemometric techniques can be used to resolve