The cloning, expression and purification of full-length HYPK has been previously described [4] . N-terminal (residues 1 - 60) and C-terminal (residues 80 - 129) regions of HYPK were cloned into pET 28a+ (Novagen). HYPK protein fragments were expressed in Escherichia coli BL21 (DE3) cells by IPTG induction (Promega). Cells were resuspended in 50 mM Tris-Cl pH 8.0, 300 mM NaCl, 1 mM PMSF, 2 mM b-mercaptoethanol and lysed by sonication on ice. The proteins were purified by Ni-NTA affinity chromatography (Qiagen). To dilute the imidazole, samples were dialyzed in 20 mM Tris-Cl, pH 7.5. Protein concentrations were determined by Bradford assay.

A protein concentration of 70 μM was used for CD experiments. To monitor unfolding proteins were incubated with 0.5, 1.0, 3.0 and 5.0 M NaClO₄ for 2 h and far-UV CD spectra were recorded on a CD spectrometer (BioLogic Science Instruments) using a rectangular quartz cell (path length; 0.1 mm). Measurements were taken between 190 nm and 250 nm at a scan rate of 30 nm/min. A total of five scans were averaged to obtain each spectrum and they were baseline subtracted for buffer (30 mM Tris, pH 8.0; 200 mM NaCl). Percent helicity values were calculated in each case using a previously described equation [16]. The plot for mean residual ellipticity (MRE) versus scan wavelength (nm) was obtained using Origin 8.0. Data smoothing was performed by adjacent averaging with a 25 point window.

HYPK has a wide fluorescence emission peak encompassing 330 - 350 nm compared to L-tyrosine which has a distinct peak at 303 nm (Figure 1(a)). To rule out the possibility of tryptophan contamination we treated the purified protein with the tryptophan specific reagents Nbromosuccinidimide and 2-Hydro-5-nitrobenzyl bromide [17]. There was no marked change in absorption and fluorescence spectra of HYPK with these treatments suggesting the absence of contamination (data not shown). Despite the fact that most of the tyrosine residues in proteins have a 10-fold lower fluorescence yield compared to free tyrosine [18,19] , the intensity of HYPK fluorescence was about 34 times less than that of L-tyrosine (Figure 1(a)). The absolute quantum yield of intrinsic fluorescence of HYPK was only 0.00414 [18-20] . Hence, the intrinsic fluorescence of HYPK was atypical not only in its emission spectra but also in its fluorescence yield. These anomalies indicated a possible microenvironment of secondary and tertiary structures in the vicinity of the lone tyrosine in HYPK.

Previous work has shown that HYPK undergoes a conformational compaction leading to aggregation in increasing Ca²⁺ concentrations. The aggregation reaction is concomitant with a reduction in fluorescence intensity [4]. To monitor conformational changes in the region of Y⁴⁹ we determined the accessibility of the tyrosine residue with increasing concentration of Ca²⁺ by acrylamide quenching. Considering the low fluorescence quantum yield of HYPK, we used a protein concentration of 70 μM [4]. To rule out any possible effect of residual Ca²⁺ in experimental buffer, all fluorescence experiments with increasing Ca²⁺ concentrations were performed after dialyzing the protein against Tris-Cl buffer (pH 7) containing 5 mM of EGTA. We determined the fluorescence intensities of HYPK at 340 nm at variable Ca²⁺ concentrations and increasing acrylamide concentrations until no aggregation of the protein was observed. The K_sv (Stern Volmer constant) values were calculated and plotted according to the equation of collisional quenching by Stern-Volmer (Eq.1) (Figure 1(b)) [21]. Only a marginal variation in the Stern-Volmer constant (K_SV) values was observed (Figure 1(c)), suggesting no appreciable alterations in the quenching pattern of HYPK intrinsic fluorescence with increasing Ca²⁺ concentrations. Despite the compaction induced, variation in Ca²⁺ concentrations did not affect the accessibility of Y⁴⁹ of HYPK by acrylamide. The precise effect of Ca²⁺ was therefore at the global level, altering the loose ensemble of (pre-) molten globules, without significantly influencing the interior formation of the secondary structure.

Increasing concentrations of GdmCl induced a decrease in fluorescence intensity of HYPK at 340 nm along with the appearance of a tyrosine emission spectrum with a maximum at 303 nm (Figure 2(a)). However, the emission at 340 nm still existed even in 6 M GdmCl. We determined the fluorescence emission of HYPK in different GdmCl concentrations and in the presence of increasing concentrations of acrylamide. As there were two different fluorophores (at 303 and 340 nm) with differential accessibilities to acrylamide, the fluorescence quenching data were analyzed by the modified SternVolmer equation (Figures 2(b)-(c)) [15]. The emission at 303 nm was quenched by acrylamide, but the emission at 340 nm was not. This result indicated that in the native condition of the protein, tyrosine fluorescence was completely quenched by intra-molecular interactions resulting in emission at 340 nm. Upon GdmCl-induced unfolding, intra-molecular quenching was lost and the tyrosine became accessible to acrylamide. However the fluorophore, emitting at 340 nm, was not accessible to acrylamide as only minimal quenching was observed.

NaClO₄ is a protein denaturant that differs from guanidine hydrochloride as it leaves helical structures unchanged [22,23]. In addition, NaClO₄ induces the formation of compact molten globule states by shielding charge repulsion effects within the molecule stabilizing the unfolded state [23]. In our study, treatment with upto 5 M NaClO₄ did not result in any shift in HYPK emission maxima; however, there was a reduction in the fluorescence intensity (Figures 3(a)-(b)). As HYPK is an IUP, it is possible that a charge effect results in conformational compaction of the molecule upon addition of sodium perchlorate. This result suggests that the lone tyrosine residue of HYPK is part of an α-helical region which the perchlorate ions cannot destabilize.

We measured the effect of pH on HYPK intrinsic fluorescence. At neutral and slightly alkaline pH, the emission peak was approximately 340 nm (Figure 3(c)). In

alkaline solutions (pH ≥ 10), the emission peak was red-shifted to ~345 - 350 nm with a reduction in intensity. Ionization of Y⁴⁹ in alkaline solution was similar to that of free tyrosine which has a weak emission at 345 nm [24]. In contrast, at acidic pH, the peak was shifted to 303 nm. However, due to insoluble aggregation at pH 5 (the pI of HYPK is 5.13) no clear fluorescence data could be recorded. From these data it appears that the effect of pH on HYPK conformation is similar to GdmCl.

To obtain further insights we performed CD spectroscopy with full-length HYPK and its N-terminal and Cterminal fragments. HYPK contains very little secondary structure as reported previously [4] . In the presence of low concentrations of NaClO₄ the percent α-helicity of full-length HYPK was increased (Figure 4(a)). However, at higher NaClO₄ concentrations the percent α-helicity was again reduced. The N-terminal HYPK (which contains the tyrosine residue) had almost no secondary structure (Figure 4(b)). On the other hand, C-terminal HYPK had substantial α-helicity in the presence of increasing concentrations of NaClO₄ (Figure 4(c)). These results suggest that with increasing concentrations of NaClO₄, the initial increase in helicity in the full-length protein may be due to the C-terminal region which is then gradually compensated by the non-helical N-terminal part.

Examples of residual structures in IUPs are limited in the literature [25,26]. The anomalous fluorescence of HYPK (an IUP), provided us with an opportunity to study its conformational flexibility. Similar proteins with a single tyrosine but no tryptophans are known to have unusual spectroscopic behavior. For example, the excitation spectrum of adrenodoxin has a maxima at 295 nm [27]. The excitation maximum of HYPK was similar to L-tyrosine at ~280 nm (data not shown). Nevertheless, the emission spectrum of adrenodoxin resembled HYPK with a peak at around 330 - 340 nm but with a higher fluorescence quantum yield [27]. These observations led to the hypothesis that ground state proton transfer contributes to the unusual tyrosine fluorescence of adrenodoxin [27]. Although the emission maximum coincided with the tyrosinate emission peak and the fluorescence quantum yield of HYPK appeared to be compatible with the very low fluorescence yield of tyrosinate [24], the anomalous intrinsic fluorescence of HYPK may be assigned to an intramolecular excited state proton transfer from phenolic hydroxyl to a proton acceptor.

In our study, the various denaturants may be disrupting secondary and tertiary structures, therefore affecting the interactions between Y⁴⁹ and one or more components of the peptide chain. Normalization of the native tyrosine fluorescence to tyrosine fluorescence of HYPK unfolded with GdmCl and low pH confirms that the intrinsic fluorescence of native HYPK was solely due to an altered conformational state of Y⁴⁹. The pK_a value of tyrosine is ~10.3, therefore when HYPK is present in a high pH solution there is an increased population of ionized and

stabilized tyrosine. Loss of fluorescence intensity at higher pH therefore indicates a further compaction of ionized tyrosine possibly through local hydrogen bonding, influenced by neighboring charge(s) or microscopic dielectric constant. GdmCl had a global unfolding effect, whilst sodium perchlorate failed to perturb the tyrosine environment of HYPK. It is known that sodium perchlorate does not unfold helical structures; therefore our data strongly support the presence of a stable helical structure involving Y⁴⁹.

Fluorescence quenching by acrylamide is known to be affected by oligomerization which alters the local environment of tyrosine side-chains [26]. However, the acrylamide quenching pattern of HYPK did not change appreciably with increasing Ca²⁺. In contrast, quenching in the presence of increasing GdmCl concentrations showed an unfolding of the HYPK structure, causing maximal accessibility of the tyrosine but not the chromophore at 340 nm. This suggests that there is a dynamic equilibrium between two populations of HYPK. The presence of highly dynamic ensembles on interconverting structures is typical of IUPs in solution [28].

Secondary structure prediction suggests that the lone tyrosine of HYPK is located in a helix (Supplementary Figure 1). Experimentally, however, we did not observe any secondary structure in the N-terminal fragment (which includes Y⁴⁹). CD spectroscopy suggested α-helical conformation only in the C-terminal region. Interestinglythe helical content of the full-length protein increased in low concentrations of NaClO₄. This might be due to stabilization of the helices which are destabilized again with a further increase in NaClO₄ concentration. HYPK contains a number of acidic residues in the local environment of Y⁴⁹. In a loosely bound molten globule, intramolecular interactions between the phenolic hydroxyl groups of the tyrosine residue and a carboxyl group of the acidic residues may result in formation of transient local structures. Absence of any phenylalanine residues in the protein excluded the possibility of long-range interactions between a phenylalanine and tyrosine. Similar structural compaction of highly charged IUPs in the presence of denaturant has been observed by single molecule FRET [29] .

There is increasing evidence that IUPs are important in biological pathways. Our previous work has confirmed that IUPs are prevalent in neurodegenerative disease networks [30]. We have previously shown that HYPK interacts with Htt and reduces Htt toxicity by modulating the aggregation kinetics by a chaperone-like mechanism. The work of Arnesen et al. suggests that HYPK interacts with NatA and cotranslationally acetylates the N-terminus of Htt, preventing Htt aggregation and toxicity [31] . Together these studies suggest that HYPK participates in several protein-protein interactions, and may have multiple functionally relevant conformations, an intriguing property of IUPs. This study demonstrates that a conformation-driven microenvironment of the tyrosine residue gives rise to the anomalous spectroscopic properties of HYPK, while the rest of the structure remains highly flexible. This structurally defined region of HYPK may have a direct correlation to its function by forming intra-molecular contacts with the flexible part of the protein.