Chemistry of conjugation to gold nanoparticles affects G-protein activity differently

Background Gold nanoparticles (AuNP) are extensively used as biophysical tools in the area of medicine and technology due to their distinct properties. However, vivid understanding of the consequences of biomolecule-nanomaterial interactions is still lacking. In this context, we explore the affect of conjugation of Gαi1 subunit (of heterotrimeric G-proteins) to AuNP and examine its consequences. We consider two bio-conjugation strategies covalent and non-covalent binding. Results Affinity of the AuNP to the Gαi1 is 7.58 × 10 12 M-1. AuNP conjugated Gαi1 exhibits altered kinetics of activation, non-covalent bio-conjugates displays retarded kinetics, up to 0.88 fold when GTPγS was used as ligand, of protein activation contrary to covalent conjugates which accelerates it to ~ 5 fold. Conjugation influence intrinsic Gαi1 GTPase function in conflicting modes. Non-covalent conjugation inhibits GTPase function (decrease in activity upto 0.8 fold) whilst covalent conjugation drastically accelerates it (12 fold increase in activity). Altered basal nucleotide uptake in both types of conjugates and GTPase function in non-covalent conjugate are almost comparable except for GTPase property of covalent conjugate. The effect is despite the fact that conjugation does not change global conformation of the protein. Conclusion These findings provide clear evidence that nanoparticles, in addition to ‘passive interaction’ with protein (biomolecule), can interact “actively” with biomolecule and modify its function. This concept should be considered while engineering nanoparticle based delivery systems in medicine.


Background
Impressive developments have occurred in nanoscience technology in the past decade, despite which a detailed understanding of nanoparticle (NP) interaction at cellular, sub-cellular and biomolecule level is lagging behind [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19]. Cedervall et al. have demonstrated that binding and dissociation parameters of protein-nanoparticle complex depend on surface characteristics of nanoparticle as well as physico-chemical properties of the protein [20]. It has been demonstrated that NPs can elevate the rate of protein fibrillation potentially leading to proposals of novel mechanisms for amyloid diseases offering therapeutic opportunities for treatment [21]. Further, imaging studies provide crucial information that nano-conjugation uniformly promotes endocytosis of EGFR, influencing its compartmentalization, and the mechanism of endocytosis [22]. Thus, nano-conjugation cannot be construed as an innocuous tool but may directly alter the cellular processes at the molecular level [22]. Improved understanding of the interactions at nano-bio interface will give answers to questions concerning the effect of conjugation on protein conformation and hence its function.
Majority of the drugs target GPCR, which transduce signal by activating heterotrimeric G-protein which in turn switches on a cascade of downstream signal transduction pathway. The activation status of heterotrimeric G protein regulates the downstream cascade events. Hence, Gα i1 is a very important model protein to investigate the effects of different types of conjugation to nanoparticle. G proteins are ubiquitously expressed and despite the variety in their function and biochemical effects, their structures are very highly conserved. These properties of G proteins, additionally, make them very vital model systems for studying the effects of nanoparticles; an area that is fast gaining importance in biology and medicine.
In the present study, we investigate the effect of the different conjugation strategies on the conformation and function of G proteins. A comparative study is presented, between non-covalently and covalently bound AuNP-Gα i1 conjugates. In the non-covalent conjugate, the rate of basal nucleotide uptake was retarded in a concentration dependent manner of AuNP, whereas in the covalent conjugate, the rate was accelerated. Both types of conjugation influenced the intrinsic Gα i1 GTPase function affecting the kinetics of GTP hydrolysis in opposite modes. Noncovalent conjugation showed inhibitory effect on GTPase function whilst covalent conjugation dramatically accelerated it. We propose that the mode of interaction with nanoparticles modulate the function of the protein in the conjugate, which may alter related cellular physiological pathways. These findings provide strong evidence that nanoparticles can interact "actively" with biomolecules and modify their function.

Bio-conjugation exploiting two different approaches
In this study two linkage strategies have been used for the conjugation of AuNPs to Gα i1 . Dihydrolipoic acid (DHLA) capped AuNPs of hydrodynamic diameter6 nm was used in the entire study. Interaction studies were performed in a buffer at pH 8.0 with low ionic strength (10 mM NaCl), since conjugates exhibited a tendency to aggregate at higher ionic strength [23].
(i) N-terminal covalent conjugation using EDC chemistry: Site specific conjugation was achieved by forming a peptide bond between N-terminal primary amine of the protein and carboxylic acid groups of negatively charged AuNP utilizing 1-Ethyl-3-[3dimethylaminopropyl] carbodiimide, EDC, chemistry. Retardation of AuNP's electrophoretic mobility on agarose gel confirmed conjugation (Additional file 1: Figure S1A). Negative control (AuNP in presence of EDC without protein) also exhibited negligibly small amount of retardation in mobility due to the formation of O-acylisourea intermediate between AuNP and EDC. Changes in mobility of Au:Gα i1 complex depends on the concentration of EDC used. (ii)Non-covalent conjugation: In non-covalent conjugation AuNP capping ligand plays an important role in the bio-conjugate. In the present study, AuNP is capped with DHLA which gives an overall negative charge on its surface. The protein may interact with AuNP in a number of orientations, or many AuNP's could be attached to a given molecule of protein.
Non-covalently bound protein-NP complex was also retarded in electrophoretic mobility, compared to AuNP itself (Additional file 1: Figure S1B), confirming their conjugation. To explore further, whether the conjugation was cysteine mediated, Gα i1 sulfhydryl groups were modified using Iodoacetamide. Conjugation to AuNP was observed even with cysteine modified Gα i1 (Additional file 1: Figure S1B), as evidenced by retardation in its electrophoretic mobility. Mobility of Gα i1 -AuNP and cysteine-modified-Gα i1 -AuNP were similar, ruling out cysteine mediated interaction between AuNP and Gα i1 . To further demonstrate that non-specifically conjugated Gα i1 has free sulphydryl groups, N-(3-pyrene) maleimide (NPM) was used to check formation of fluorescent adducts with free thiol groups (Additional file 2: S2). Cysteinemodified Gα i1 -AuNP, upon treatment with NPM displayed fluorescence spectrum with peaks at 377 nm, 397 nm and 418 nm similar to Gα i1 -NPM adduct, though, with a lesser intensity. These results confirm the presence of free sulphydryl groups of Gα i1 even after non specific conjugation with AuNPs.

Quenching of Gα i1 Tryptophan fluorescence by AuNP
Tryptophan fluorescence of Gα i1 was quenched by AuNP in a dose-dependent (0.1-0.5 nM) manner ( Figure 1). No shift was observed in the λ max,em of tryptophan Gα i1 -AuNP conjugate formation, indicating that the polarity of tryptophan environment, and hence the overall protein structure did not change upon conjugation. The binding constant (K b ) and the numbers of binding sites (n) between AuNPs and Gα i1 were determined using the method described by Tedesco et al., [24] as 7.58 × 10 12 M -1 and 1.2 respectively, from the fluorescence spectral titration.

Rate of activation is differentially affected by the nature of conjugation
Fluoroaluminates activate Gα i1 -GDP by mimicking the γphosphate of GTP in its binding site. Time dependent fluorescence changes from Gα i1 upon activation by AlF 4 binding was monitored for non-covalent and covalent AuNP conjugated Gα i1 and activation rates were calculated.
(i) Non-covalent conjugation: Attenuation in the rate of activation accompanied with decrease in values of maximum plateau in fluorescence by AlF 4 were observed with to non-covalent complex of Au:Gα i1(GDP) . The effect was AuNP concentration dependent, with total loss of activity at 0.4 nM AuNP for 200 nM Gα i1 ( Figure 2A, Table 1).
(ii) Covalent conjugation: Conversely, covalent conjugation at the N-termini of the protein caused enhancement in the rate of AlF 4mediated activation, 3.2 fold in comparison to unconjugated protein ( Figure 2B, Table 1).
Both non-covalently and covalently bound AuNP did not perturb the characteristic feature of Gα i1 to bind GDP nucleotide and its behaviour to undergo activationdependent changes induced by transition state mimetic, AlF 4 -. Non-covalent and covalent conjugation, modulated kinetics of AlF 4 induced activation of Gα i1 in contrasting manner. The rate of activation by AlF 4 is much faster in case of covalently conjugated protein and the peak fluorescence of active protein was comparable with respect to unconjugated protein ( Figure 2).

Conjugation does not affect the secondary structure of the protein
Far-UV Circular Dichroism (CD) spectra were recorded to monitor secondary structural features of the protein. Noncovalent or covalent conjugation with AuNP did not cause changes in secondary structure of the protein suggesting the global structure of complex of Au:Gα i1(GDP) to be intact (Additional file 3: Figure S3). These findings clearly indicate that the conjugation of AuNP changes the activity of Gα i1 without affecting the conformation of the protein.

Mode of interaction between Gα i1 and AuNP alter the kinetics of basal GTPγS binding
Next, we investigated whether functionalization of Gα i1 with AuNP affects the basal nucleotide exchange rate, the Hepes-Na (pH 8.0)) were mixed with protein and incubated for 10 minutes. Tryptophan emission at 340 nm was monitored by exciting the sample at 295 nm with continuous stirring. 2 mM NaF and subsequently followed by 20 μM AlCl 3 was added to the reaction and relative fluorescence was monitored as a function of time. All the measurements were performed at 25°C. Non-covalently conjugated AuNP-Gα i1 displayed deaccelerated rates of basal AlF 4binding to Gα i1 -GDP. Non covalent conjugates decreased AlF 4binding upto 0.08 fold. On the contrary N-terminal covalent conjugation caused 3.2 fold increase in rate of AlF 4binding. The Plateau fluorescence intensity of covalently conjugated AuNP-Gα i1 was comparable to only Gαi1, whereas non-covalent conjugation displayed decrease in plateau fluorescence in a concentration dependent manner.
in vivo activity of the protein. Nucleotide exchange (GDP to GTPγS) by Gα i1 , upon covalent and non-covalent conjugation of AuNP, was monitored by measuring the enhancement in intrinsic Trp fluorescence. Changes in fluorescence were monitored as a function of time after addition of GTPγS ( Figure 3, Table 2). Non-covalent conjugation led to a drop in the basal rate of GTPγS uptake, while covalent conjugation caused an increase of~5 fold in the rate of GTPγS uptake. Rate of GTPγS uptake by both types of AuNP:Gα i1(GDP) complexes corroborated GDP/AlF 4 activation data. Both results provide evidence for dependence of functional behaviour of conjugates on the nature of interaction between AuNP and Gα i1 , as the conjugates preserved the native conformation confirmed by far UV CD analysis.

AuNP conjugation modulates Gα i1 intrinsic GTPase activity
We used an extrinsic fluorescent probe, N' -Methylan thraniloyl (mant)-GTP (mGTP) in order to quantitatively study effect of AuNP conjugation on binding and release of nucleotide (GTP/GDP) and monitor Gα i1 activation. Fluorescence resonance energy transfer (FRET) was monitored as a function of time by exciting the intrinsic tryptophan fluorescence at 295 nm and measuring the mGTP fluorescence at 448 nm. 200 nM Gα i1 was titrated with several concentration of mGTP (100 nM to 800 nM). mGTP fluorescence increased upon addition to Gα i1 and then decayed at a slower rate, confirming mGTP hydrolysis to mGDP. Further, addition of 10 μM GTPγS decreased the mGDP fluoresecence rapidly. To obtain the corrected mGTP uptake and hydrolysis, fluorescence remaining after GTPγS addition was subtracted.
Non-covalent conjugation of Gα i1 with AuNP retarded the basal exchange rate of mGTP to protein bound GDP, up to 0.4 fold, in a concentration dependent manner of AuNP ( Figure 4A, Table 3) Altered Gα i1 -NP basal nucleotide uptake convincingly demonstrates AuNP influence of the mode of binding. Further, to investigate the effect of AuNP conjugation on Gα i1 GTPase function mGTP hydrolysis kinetics was monitored. Significant decrease in mGTP hydrolysis rate ( Figure 5A) was observed (Table 4), suggesting the non-covalent binding with AuNP has an inhibiting effect on intrinsic GTPase property of Gα i1 . Covalent conjugation of Gα i1 with AuNP resulted in 12 fold increase in GTPase activity ( Figure 4B). This indicates that covalent conjugation of AuNP to Gα i1 has an accelerating effect on intrinsic GTPase function of Gα i1 .

Discussion
In the present study, we address the dependence on the chemistry of conjugation towards alteration in the kinetics of activation of Gα i1 . There have been contradicting observations regarding the benefit of using a nanoparticle in medicine and biochemistry [25][26][27]. In our view, understanding the physico-chemical basis of how an engineered nanoparticle modulates biological processes requires the study of nanoparticle-biomolecule binding and its effects on biomolecule functionality. Recent studies have emphasized that properties like size, shape, surface modification, and charge of nanoparticles can profoundly affect the interaction between NPs and biomolecules. Guanine nucleotide binding proteins (G-protein) play a vital role in the physiology of a cell. Structure-function relationship of both, the monomeric and the heterotrimeric G proteins are well understood, their crystal structures helping elucidate their mode of action and the biochemical function. Heterotrimeric G-proteins are activated by agonist-stimulated G Protein-Coupled Receptors (GPCRs) that catalyze the exchange of GTP for GDP on G protein α-subunits and relay extracellular signal to intracellular signalling pathways [28]. We chose to use G proteins as model proteins to better understand the effect of AuNP binding to proteins and the biological effect they elucidate.
Here, we exploit two modes of conjugation between Gα i1 and AuNP, covalent and non-covalent. Effect of AuNP conjugation to Gα i1 was examined by monitoring steady-state Trp-fluorescence from the protein. AuNP interacts with Gα i1 with a binding constant (K b ) of 7.58 × 10 12 M -1 . Strong tryptophan fluorescence quenching of Gα i1 was observed with increasing concentration of AuNP. Fluorescence quenching could be explained by efficient energy transfer between AuNP and Gα i1 tryptophan residues. No shift in emission wavelength was noticed, suggesting no change in polarity around tryptophan residues on addition of AuNPs.
Cysteine modifications did not alter the AuNP conjugation and, vice versa, conjugation did not lead to the unavailability of the thiol groups of cysteine for the modifying reagents, therefore leading us to conclude that the 10 cysteine residues (of Gα i1 ) do not interact with AuNP via thiol-Au linkage chemistry. Our finding is in agreement with a previous study which demonstrates that cysteine residue at the end of a C-terminus of protein was much more reactive toward a gold cluster than a cysteine residue introduced in middle region of protein [29]. This study concludes that non-specific AuNP interaction is not protein-sulphydryl mediated even though cysteine residues are present on Gα i1 surface.  A number of studies have shown that nanoparticle protein conjugates undergo conformational changes and result in unfolding of protein [30][31][32]. For biochemical applications of NP-protein conjugates, it is crucial that labelling does not modify the protein structure. Interestingly, both types of conjugates of Gα i1 retain their secondary structure as evident from far UV-CD spectra profile for Gα i1 and the conjugates (Additional file 3: Figure S3).
We next investigated whether the bioconjugated Gα i1 was functional and active. We report here functional activity of both the covalent and non-covalent AuNP conjugated Gα i1 . Time dependent fluorescence measurement using intrinsic tryptophan and extrinsic MANT moiety fluorescence with hydrolyzable and non-hydrolyzable nucleotides were assayed. Detailed kinetics based functional studies for both non-covalent and covalent AuNP-Gα i1 conjugates have provided important insights: (i) reduced rate of activation by AlF 4-, GTPγS and mant-GTP were observed as a consequence of non-covalent interaction of AuNP. (ii) N-terminal covalent probing led to enhanced rate of nucleotide uptake "activity" of Gα i1 .
To further probe the influence of AuNP conjugation on intrinsic Gα i1 GTPase function, kinetics of bound GTP hydrolysis was examined. As demonstrated in Figure 5A, non-covalent conjugation does not affect the GTPase function while covalent conjugation dramatically accelerated it. From our studies, it may be hypothesized that the functional property of conjugated protein are governed by the contribution of type of molecular interaction, between the nanomaterial and biomolecule ( Figure 6). This has strong implication that nanoparticles can impair the cell function when it enters into biological fluid depending on the extent and format of presentation of the signalling protein to the nanoparticles. Thus, these finding presented here need   to be considered carefully before using engineered nanoparticles for medical application.

Conclusion
In summary, we here report two different bioconjugation strategies, non-covalent and covalent attachment, of Gα i1 to 6 nm DHLA capped AuNP. No effect of change in protein conformation was observed despite the presence of negatively charged capping ligand, DHLA. Non-covalent bioconjugation caused decrease in "activity" of Gα i1 in terms of decelerated rate of nucleotide exchange and inhibited GTPase activity. N-terminal covalent probing of AuNP modulate the active state of Gα i1 state, as displayed by enhanced rate of nucleotide exchange and stimulated GTPase function. These results (extraordinary increase in the Gα i1 GTPase property) have ramification in understanding the probable molecular basis of gold to cure many diseases when used either in powder form (in ayurvedic treatment) or as colloidal gold in modern medicine (e.g., in arthritis).

Materials
All the chemicals used were purchased from Sigma-Aldrich, USA.

Size and optical characterization of AuNP
Transmission electron microscopy was used to visualise the shape and to determine size distribution of AuNPs (Additional file 5: Figure S5). TEM images were obtained using JEOL 3010, operating at 300 kV accelerating voltage. The average size distribution was determined by using the image analysis software, Image J. The UV-vis absorption spectra were recorded on a Jasco V-660 UV-vis spectrometer at room temperature with 1 cm path length cuvette. Spectra were obtained with a band width of 1.0 nm and a scan rate of 40 nm/minutes (Additional file 6: Figure S6). Toluene was used as reference.

Expression and purification of Gα i1 protein
The DNA fragment containing the WT rat Gα i1 subunit, cloned into the pET28b expression vector was used to transform BL21 (DE3) cells to express an N-terminal hexa-His-tag-WTGαi1 protein in the presence of kanamycin (100 μg/mL) and purified (Additional file 7: S7). The eluted protein fractions with the maximum protein content were estimated by Lowry's method [34] and purity checked by SDS-PAGE. The average yield for WT Gα i1 was 10 mg/L of culture.

Fluorescence spectroscopy Steady state fluoresecence
Steady state intrinsic tryptophan fluorescence of Gα i1 in inactive form was recorded on HORIBA Jobin Yvon fluorolog spectrometer with excitation light of 295 nm (excitatoin and emission slit width of 5 nm) at 25°C. In a 3 mL quartz cuvette, 400 nM Gα i1 [in 5 mM Hepes-Na (pH 8.0), 10 mM NaCl, 0.5 mM MgCl 2 and 1 μM GDP] was taken and titrated with AuNPs. In all cases blank spectra (buffer containing only AuNP) were subtracted from the protein spectra.

Fluorescence-based kinetic assays
Time-based fluorescence activity measurements were performed on a Jasco FP-6500 Spectrofluorometer at 25°C. In a 3 mL cuvette, 400 nM Gα i1 [in 5 mM Hepes (pH 8.0), 10 mM NaCl, 0.5 mM MgCl 2 and 1 μM GDP] was taken. 16 μM GTPγS was added to the protein and the relative increase in intrinsic fluorescence (λex = 295 nm, λem = 340 nm) was measured as a function of time. Similar measurments were performed for AuNP conjugated Gα i1 . GTPγS exchange rates were determined as described elsewhere [35]. FRET was monitored by exciting the intrinsic Trp fluorescence at 295 nm and measuring the mant-GTP fluorescence at 448 nm. In all cases, blank spectra containing buffer alone were subtracted from the final spectra.
Circular dichroism (CD) spectroscopy CD measurements were made on a JASCO model J-715 spectropolarimeter. Far-UV-CD spectra were recorded in 1 cm path length cuvette from 200 to 260 nm; each spectrum was the average of 5 scans. Spectra were recorded with the final protein concentration of 100 μg/mL.