Biophysical and structural studies reveal marginal stability of Acyl ACP Reductase: a crucial hydrocarbon biosynthetic enzyme

Background : Acyl-ACP reductase (AAR) is one of the two key cyanobacterial enzymes along with aldehyde deformylating oxygenase (ADO) involved in synthesis of long-chain alkanes, a drop-in biofuel. The enzyme is prone to aggregation when expressed in E. coli , leading to varying alkane levels. Intriguingly, the structural characterization remains largely elusive as AAR alone failed to form stable crystals, possibly due to a number of intrinsically flexible random regions. The present work attempts to fill a gap in the literature by investigating the crucial structural aspects of AAR protein associated with its stability and folding. Results : The AAR protein was recombinant expressed in E. coli and purified by metal affinity and gel filtration chromatography. Characterization by dynamic light scattering experiment revealed that recombinantly expressed AAR in E. coli existed in multiple-sized protein particles in the range of 36.4 to 51.6 nm. Intact mass spectrometry revealed that recombinant AAR was heterogenous due to diverse lipidation and de-lipidation resulted in a single mass peak of 40296.87 Da as predicted. Interestingly, while thermal- and urea-based denaturation of AAR showed 2-state unfolding transition in circular dichroism and intrinsic fluorescent spectroscopy, the unfolding process of AAR was a 3-state pathway in GdnHCl solution. Lower concentration of GdnHCl appeared to be stabilizing the protein, suggesting that the protein milieu plays a significant role in dictating it’s folding. Standard free energy (∆G H2ONU ) of ~4.5 kcal/mol for steady-state unfolding of AAR indicated borderline stability of the protein. Molecular dynamics simulation conducted on AAR structure in presence of KCl, an ionic solvent with similar properties as GdnHCl at lower concentrations, suggested that KCl mediates structural stabilization especially at the concentration of 375 mM, and thus was responsible for enhancing its activity. KCl presence also resulted in regional alteration towards the binding site of its neighbouring pathway enzyme, ADO, thus paving the way for coordinated catalysis. Conclusion : Based on these evidences, we propose that the marginal stability of AAR are plausible contributing reasons for aggregation propensity and hence low catalytic activity of the enzyme when expressed in E. coli for biofuel production. Our results show path for building superior biocatalyst for higher biofuel production.


Background
The emerging unavailability of fossil fuels, heightened levels of pollutants, and greenhouse gas emissions have directed attention towards the development of safe and renewable energy sources [1,2]. Biofuels have the potential to provide a sustainable cost-effective alternative for petroleumbased hydrocarbons. High energy density and compatibility of the long-chain alk(a/e)ne with the existing engines and infrastructure furthers its demand as an effective and ideal candidate for dropin biofuels [2,3].
AAR is a key enzyme of the pathway and is prone to aggregation and the majority of protein forms inclusion bodies when expressed in E. coli [14]. Interestingly, the alkane level varies among different cyanobacteria strains with variable soluble AAR levels [14]. AAR has been reported recently to contain three flexible regions possessing intrinsic flexibility in the N-terminal and middomain of the structure (Fig. 1A) [15]. AAR consists of three domains: the N-terminal domain (NTD, residues 1-130), the middle domain (mid-domain, residues 131-264) with a conserved dinucleotide recognition loop for NADPH binding, and the C-terminal domain (CTD, residues 265-341) [15]. Residue C294, which is strictly conserved in AAR homologs, is located at the center of the AAR molecule [15].
Studies have been carried out exploring the functional characterization of AAR and gaining structural insights into the mechanistic aspects of AAR-ADO function [10,12,14,15]. However, gaining deep insights into the structural aspects of the AAR protein remains largely elusive. Proper folding, conformation, and stability are very crucial factors to form functionally active AAR protein. Reports are available where ADO, the consecutive enzyme in the pathway, has exhibited remarkable enhanced activity upon improvement in the stability of the protein [2,16].
Therefore, it is imperative to understand the stability and folding of the protein, given the immense biosynthetic potential of AAR in industries for hydrocarbon production. However, so far there are no reports available regarding the conformational and thermodynamics stability characterization of AAR, which are among the major goals for industrially relevant proteins that are likely to be exposed to extreme conditions such as high temperature, presence of co-solvents, low pH, etc.
The stability of a native protein is a function of surrounding environmental variables such as pH, temperature, ionic strength, and solvent composition as they perturb various intramolecular bonds, responsible for stability and integration of the protein. In addition, only perfectly folded proteins survive for long term in biologically crowded environment and interact selectively with their natural ligands. Failure to fold into the intended conformation usually produces inactive and misfolded proteins with undesirable properties. For proteins of biotechnological significance like AAR, understanding of environmental factors that dictate stability and folding will be important for better production of the functional recombinant enzyme at industrial scale. 6 In the present study, we purified the heterologously expressed AAR in E. coli and performed detailed in vitro physico-chemical characterization of AAR to gain insights into elucidating the thermodynamic parameters and conformational stability of the protein. We studied the equilibrium unfolding pathway of AAR by carrying out chemical denaturant-, and heat-mediated unfolding experiments, and the unfolding data was used for the calculation of thermodynamic stability parameters associated with the processes. The changes in the secondary and tertiary structural folds were measured by CD and intrinsic fluorescence-based spectroscopic methods. Unfolding transition behaviour along with the parameters like melting temperature (Tm), mid-point of the unfolding transition (Dm) and the standard free energy change parameters (∆G H 2 O NU) were further elucidated for the AAR protein. These data were indicative of the marginal stability of the AAR protein which is a plausible reason for the aggregation propensity, low turnover number and hence low catalytic activity of the enzyme when expressed in E. coli. Furthermore, molecular dynamics (MD) simulation conducted on AAR structure enabled us to understand the positive effect of ionic solvent KCl on protein structure stabilization.

Expression, purification and functional characterization of AAR
For overexpressing AAR, the codon-optimized aar gene was cloned into the bacterial expression vector pQE30 [8], transformed into M15 strain of E. coli, and induced in liquid culture using 1 mM IPTG. The induced culture was shifted to 18 o C and cells were harvested after 16 h of postinduction. SDS-PAGE analysis of the induced culture pellet revealed an intense band around 38 kDa, which was consistent with the calculated molecular mass of AAR and was absent in the uninduced cells (Fig. 1B). 7 The purification of 6-His-tagged AAR was carried out in a two-step manner (Fig. 1B). The first step of purification involved Ni-NTA based affinity chromatography, where major contaminant proteins in the supernatant were removed. The employment of the second step, i.e. gel filtration chromatography, helped in the removal of the residual non-AAR proteins (Fig. S1) and the purified AAR fractions were pooled, concentrated and analyzed by SDS-PAGE (Fig. 1B). The results suggest that AAR has been successfully purified by the two steps of chromatographic purification.
Functional AAR is known to exhibit reductase activity catalyzing the acyl-CoA/ACP substrate to its corresponding aldehyde. We checked the activity of the purified AAR using palmitoyl-CoA and NADPH as substrates and analyzed the products of the reaction at different time intervals by GC-MS (Fig. 2). The maximum product concentration peak with retention time 13.9 min, corresponding to 4.7 µg/mL hexadecanal, was achieved at 4 h of incubation period ( Fig. 2A and   2B), as analyzed by GC-MS profile (Fig. 2C). The result suggests that the purified AAR is a functionally active protein.

Determination of disperse nature of purified AAR
We realized from the GFC profile that AAR eluted at wide elution volume rather than a sharp peak ( Fig. S1). To understand the nature of AAR eluted at different retention time, we determined its dispersity via dynamic light scattering (DLS) experiment. AAR GFC elutes were divided into four different fractions, namely a, b, c, d (Fig. S1), pooled and concentrated to ~1 mg/mL for performing DLS. It was observed that a, b, and c fractions were monodisperse with varying hydrodynamic diameter of 51.6 nm, 45.3 nm, 36.4 nm, respectively (Fig. 3). On the other hand, Fraction d was polydisperse in nature. Fraction a, b, and c corresponds to the molecular weight of AAR when analyzed on SDS-PAGE (Fig. S1) and exhibits enzymatic activity (Fig. 2). Together when pooled, fractions a, b, and c corresponds to a polydisperse state of solution (Fig. S2). It is 8 indicative of multiple-sized protein particles that exist in AAR purified solution as depicted from the GFC profile of AAR as well (Fig. S1), thus corresponding to different states of AAR.

Analysis of intact nature of AAR using LC-Mass Spectrometry
Native protein intact mass determination of purified AAR was carried out using ESI-mass spectrometry coupled to UHPLC for the upfront separation system. Fig. 4A represents the deconvoluted mass spectrum of AAR derived from the full ESI-MS m/z scan (inset). Purified AAR appeared to adopt heterogeneous states in the solution with mass ranging from 20 to 100 kDa (Fig.   4A), which was in agreement with the DLS data (Fig. 3). Interestingly, when purified AAR was subjected to de-lipidation using Shimadzu MAYI-ODS column prior to LC-ESI MS analysis, it resulted in a single mass peak of 40296.87 Da (Fig 4B). The mass corresponded to the single monomeric molecular weight of AAR and no heterogeneous states could be observed in the deconvoluted spectra (Fig. 4B) as compared to the un-treated AAR (Fig. 4A). This result suggested that some lipid molecules bind to AAR during its heterologous expression in E. coli.
To further explore the nature of lipid molecule bound to AAR, we carried out the LC-MS analysis of the lipid extract from the ODS column (Fig. S3). The characterization of the lipids binding to AAR protein was qualitative in nature by using a non-targeted profiling method. LC-MS fragmentation data was searched across the database LipidSearch TM (version 4.1.16, Thermo  Table   S1. The most prominent of these were fatty acids, hydroxy fatty acids, and prenol lipid. Determination of structural integrity of AAR 9 The secondary structure of AAR was determined by monitoring far-UV CD spectra from wavelength 200 to 250 nm as shown in Fig. 5A. The CD spectra were used for the calculation of the percentage of the helix, sheets, turns, and random structure for AAR protein. The percentage of helix, sheet, turns, and random structures were approximately 39.1%, 22.1%, 9.7%, and 28.1%, respectively, by using Yang's reference software. The result was in agreement with the reported secondary structure composition of AAR [15] indicating that the purified protein has retained its native secondary structural folds. Intrinsic emission fluorescence of the aromatic acid residues in a protein was exploited for determining nature of the tertiary structural folds of the AAR protein by fluorescence spectroscopy. AAR comprises six tyrosine and six tryptophan residues. The protein was excited at 280 nm and 295 nm to obtain intrinsic 'Tyr+Trp' and intrinsic 'Trp' fluorescence spectra, respectively ( Fig. 5B and 5C). The emission maximum wavelength (λmax) of AAR when excited at 280 nm and 295 nm spectra was 337 and 340 nm, respectively ( Fig. 5B and   5C). The λmax values correspond to the emission of Trp residue when not exposed to the polar solvent environment, indicative of buried residues in a hydrophobic environment which is the characteristic of a folded protein. These results, therefore, suggested that purified AAR existed in its correctly folded native form.

Thermal unfolding and refolding studies of AAR
The surrounding environment, to which a protein is being subjected to, plays a crucial role in dictating its folding and stability in the cellular condition. Exposure of protein to unfavorable conditions accounts for its misfolding or unfolding leading to aggregation. To understand the effect of one of the critical environmental variables like temperature on the structural integrity of the AAR protein, we have carried out thermal-induced unfolding and refolding studies.
Study of thermal unfolding and refolding of AAR by CD spectroscopy 10 Thermal unfolding and refolding of AAR was probed by CD spectroscopy. Fig. 6A represents changes in the CD spectra of AAR when subjected to different temperatures corresponding to the changes in the secondary structure of the protein. The unfolding transition of AAR was monitored at 222 nm upon subjecting to temperature range varying from 20 o C to 90 o C (Fig. 6A). With AAR being subjected to an end denaturation temperature of 90 o C, the loss of secondary structure could be observed. A typical sigmoidal curve was obtained with increasing temperature indicative of almost complete unfolding of protein and loss of its secondary structure (Fig. 6B). The curve also suggested that the secondary structure of AAR remained intact up to 35 o C, with eventual conformational change occurring till 60 o C followed by the continuous loss of secondary structure at higher temperatures. The melting temperature of AAR was determined by fitting the thermal denaturation transition curve to a two-state model and was found to be 43.3 ± 0.9 o C ( Table 1).
The experimental data fitting coincided with the theoretical data fitting line for the two-state (Fig.   6B). The 90 o C heat-denatured protein when cooled gradually from 90 o C to 20 o C was unable to attain the folded native conformation (Fig 6B). This shows the process is irreversible in nature as depicted.

Thermal unfolding and refolding of AAR probed by intrinsic fluorescence spectroscopy
Changes in the tertiary structure of AAR with increasing temperature was also monitored by recording changes in intrinsic tryptophan fluorescence intensity with continuous scan from 20-90 o C ( Fig. 6C and 6D). A continuous decrease in the fluorescence intensity could be observed with increasing temperature (Fig. 6C). The relative fluorescence intensity corresponding to the respective temperature was plotted. The tertiary structure of AAR was retained until 30 o C as no change in the emission maximum was observed (Fig. 6D). This was followed by a gradual redshift in the λmax from 340 to 344 nm up to 60 o C indicating the conformational changes in the 11 protein and further leading to denaturation at higher temperatures (Fig. 6C). Similar to CD, fluorescence data was also fitted to a two-state model and fitting parameters obtained were useful in determination of the melting temperature of AAR to be 42.8 ± 2.1 o C (Table 1).

Denaturant-mediated equilibrium unfolding of AAR
To comprehend the nature of interactions mainly responsible for stabilizing the AAR protein, effect of different solvent composition on the structural integrity of the protein was studied via chemical denaturant-mediated equilibrium unfolding studies. AAR was subjected to two different denaturants exhibiting different chemical characteristics, namely urea and GdnHCl, and the effect on the conformation and stability of the protein was studied.

Urea-mediated equilibrium unfolding of AAR probed by CD spectroscopy
The changes in ellipticity at 210-250 nm were measured at different urea concentrations from (0-8 M) (Fig. 7A). The MRE (molar residue ellipticity) values calculated at 222 nm for each sample were plotted against the respective urea concentration from 0 to 8 M to obtain an unfolding transition curve (Fig. 7B). An increased negative MRE value at lower urea concentration (0.1-2.5 M) as compared to native protein, suggests exposure of secondary structure from the compact 3-D structure on mild treatment with a denaturant. Beyond 2.5 M concentration, a decrease in the value of negative MRE suggests a continuous loss of secondary structure till 6.0 M urea concentration. Eventually, at 6.0 M urea concentration, the maximum unfolded fraction of protein was obtained which indicated a complete loss of the secondary structure of the protein ( Fig. 7A and 7B). The unfolding transition curve was fitted into the two-state transition model (Equation 6), and the Dm value (denaturant concentration at which 50% of the protein is unfolded) was calculated to be 4.3 ± 0.4 M. The standard free energy change was also calculated and found to be 4.3 ± 1.0 kcal mol -1 respectively ( Table 2). The urea-induced unfolding of AAR can be represented by the following scheme: Urea-mediated equilibrium unfolding of AAR probed by intrinsic fluorescence spectroscopy  7D). The Dm value from the model was calculated to be 4.7 ± 0.2 M and the standard free energy change was calculated to be 5.4 ± 0.9 kcal mol -1 ( Table 2). The results obtained here further validate the earlier-calculated parameters from the CD data.

GdnHCl-mediated equilibrium unfolding of AAR probed by CD spectroscopy
The   (Table 3).Therefore, GdnHCl-induced unfolding can be represented by the following scheme:

GdnHCl-mediated equilibrium unfolding of AAR probed by intrinsic fluorescence spectroscopy
The AAR unfolding transition was monitored at different GdnHCl concentrations through the changes of fluorescence intensity and the emission maximum wavelength (λmax).

Computational evaluation of AAR stability and interactions in presence of KCl
KCl, an ionic solvent with similar properties as GdnHCl at lower concentrations, has been reported to enhance AAR activity significantly [11,15] After an extensive biophysical characterization of AAR exhibiting different unfolding properties in presence of different solvent environments, we wanted to understand the effect of KCl using computational modelling and simulation methods.

Discussion
Schirmer group in 2010 had identified two enzymes of the alkane biosynthetic enzyme pathway [10]. Since then, extensive structural characterization of the terminal ADO protein and its variants had been carried out [2]. Surprisingly for AAR, the structural aspects still remain largely obscure 18 generating a wide gap in the literature. The matter becomes crucial to be investigated considering the colossal biotechnological relevance of the protein and diminishingly existing structural information.
In the present communication, we have attempted to isolate and carry out a detailed characterization of the AAR protein. The AAR protein from Synechococcus elongatus PCC7942 had been reported to exhibit the maximum activity among other cyanobacterial AARs [14]. We studied the expression of this protein in E. coli since the maximum achieved hydrocarbon production to date has been reported in this host [8]. However, AAR, a water-soluble cyanobacterial enzyme, unlike ADO is prone to aggregation when expressed recombinantly in the heterologous host. The present work has successfully attempted to investigate the reasons associated with the structural discrepancies exhibited by AAR and shortlisted the probable bottlenecks associated with its aggregation which could be targeted and resolved for achieving improved hydrocarbon production.
AAR protein has been expressed and successfully purified from the E. coli host (Fig. 1). The enzyme has exhibited reductase activity thus confirming that the purified AAR is a functionally active protein (Fig. 2). Interestingly, GFC elution profile of AAR does not correspond to a monomeric fraction, instead, a broadly distributed peak-like pattern has been observed indicative of the co-existence of multiple-sized states of the protein in solution (Fig. S1), which was also previously reported by Kudo et al. group [18]. The presence of other protein contaminants was ruled out as only a single band was observed in SDS-PAGE corresponding to the monomeric molecular weight of the AAR protein ( Fig. 1 and Fig. S1). The polydispersity of the AAR purified fraction was further established by DLS confirming with varied-sized protein particles in the 19 solution (Fig. 3). Although it is difficult to quantify the oligomeric state, the data altogether offers an indication of size spread and relative contributions.
Additionally, mass spectrometry-based intact mass analysis of purified AAR exhibited irregularly distributed multiple states of AAR (Fig. 4a), further confirming the aforementioned results of GFC and DLS data. Excitingly, when the AAR purified fraction was subjected to de-lipidation, a single monomeric state of the protein dominated in the solution (Fig. 4b).
The data acquired is intriguing as it suggests fatty acid-binding induces heterogenous multiple-sized states of the AAR protein.
Reports are available where fatty acid-binding to a protein induces the oligomerization which can lead to aggregation of the protein [19]. We propose that this could be one of the contributing reasons for enhanced aggregation propensity of the otherwise water-soluble AAR protein [9] when expressed in the E. coli cells for hydrocarbon production. The current study accommodates comprehensive in vitro biophysical characterization of AAR to gain insights into the structural and stability aspects of the protein. It is possible that protein stabilities could be different inside cells than in vitro due to protein-protein interaction and other factors. However, few key experimental studies show that protein stability in cells is approximately the same as deduced in vitro [20][21][22].
We have used far-UV CD, and intrinsic fluorescence spectroscopy to know about the secondary and tertiary structural folds of AAR protein. The calculated % of secondary structure from far-UV CD spectra were ~40% helix, ~22% sheets, ~10% turns, and ~28% random coil, which suggested a properly folded secondary structure (Fig. 5) and was in concordance with the recently published structure of AAR [15]. The emission fluorescence spectra of native AAR showed an emission maximum of 340 nm (Fig. 5). The λmax values obtained (340 nm) corresponded to the emission of six Trp residues and was indicative of slightly exposed buried residues in a hydrophobic environment when not entirely exposed to the polar solvent, which is the 20 characteristic of a folded protein. The results are therefore suggestive of purified AAR to exist in its folded native form.
Thermal stability is a critical property for many biotechnological applications of proteins as it implies longer life-times and frequently higher tolerance to the presence of organic co-solvents, extreme pH values, and high salt concentration or pressures [23]. In the present work, we have reported temperature-induced denaturation and renaturation studies of AAR.
AAR upon thermal denaturation, measured by CD and intrinsic fluoresce spectroscopy (Fig. 6 Interestingly, from this data, it could be inferred that the temperature range up to which AAR resists the change in its native structure is below the E. coli cultivation temperature at which the protein is usually expressed (30-37 o C), thereby undergoing thermal stress. This is suggestive that thermal stress is one of the contributing factors to the aggregation propensity of the protein when expressed in E. coli host. Loss of functional protein occurs due to irreversible degradation or misfolding or denaturation, of the otherwise water-soluble cyanobacterial enzyme, leading to aggregation in the heterologous host. Our results, therefore, indicate that there exists an immense scope to improve the thermal stability of the AAR protein so that the protein can withstand thermal stress when expressed in E. coli and end up not forming inclusion bodies.
The chemical stability of AAR protein has also been characterized by carrying out denaturantmediated equilibrium unfolding studies. The unfolding processes of AAR induced by urea and guanidine hydrochloride (GdnHCl) were investigated by spectroscopic methods. In the unfolding processes, AAR tertiary structural transition was monitored by the changes of intrinsic fluorescence emission spectra, and its secondary structural transition was measured by the changes of far-UV CD spectra.
The urea-mediated unfolding of AAR appears to follow a single unfolding transition (Fig. 7). The protein resists the change in the secondary and tertiary structure till 2.5 M urea concentration followed by a gradual loss in the structure of the protein till 6.0 M urea concentration beyond which no further loss in the structure has been observed. The redshift in the emission maximum accompanying the loss of secondary structure indicates protein to attain an unfolded state. AAR undergoes a co-operative two-state unfolding transition between N (native) and U (unfolded) state when treated with urea. The Dm (denaturant concentration at which 50% of the protein unfolded) and standard free energy was determined to be ~ 4.0 M and ~ 5.0 kcal mol -1 respectively ( Table   3).
The GdnHCl-induced equilibrium unfolding of AAR appears to follow the double sigmoidal curve indicative of two transitions, one at lower GdnHCl concentration while another at a concentration of 2.25 M GdnHCl (Fig. 8). The protein becomes fully unfolded at 3.0 M GdnHCl concentration.
Due to the presence of more than one kind of observed population in the process, the data were Although native AAR showed partial unfolding at initial GdnHCl concentration, the intermediate state formed at 2.25 M concentration seems to regain nearly native-like secondary structure (CD spectra) (Fig. 8A), similar fluorescence quantum yield and similar fluorescence emission maxima (Fluorescence spectra) (Fig. 8C). This could be attributed to the mode of the unfolding mechanism of GdnHCl considering that GdnHCl is a monovalent salt having both ionic and chaotropic effects [24,25].
Guanidine is a salt expected to exist in the fully ionized form in aqueous solutions. The presence of Gdn+ and Cl-influence the overall stability of proteins [24]. The stability of the intermediate is interpreted in terms of stabilization by generated Gdn + and Clions masking the charged moieties of protein thereby reducing or even eliminating any destabilizing electrostatic interactions which has been reported for other enzymes as well [25,26]. Interestingly, AAR has been reported to interact with ADO primarily via electrostatic interactions [15], thus retaining the exposed charged residues on its surface, which in the present case are stabilized at the lower concentration of the  [11,15]. KCl is known to exhibit ionic character similar to the low concentration of GdnHCl [27].
The MD simulation studies revealed the effect of KCl on the N-terminal domain of AAR, 23 responsible for interaction with the ADO protein, and also the diminished fluctuations in the flexible region III was observed when protein was subjected to solvent with 375 mM KCl (Fig. 9).
The acidic residues of this helix 7 of ADO are known to form strong electrostatic interactions with the basic residues from the long helical region (R73-H91), as well as R118 from flexible region II of AAR [15]. Interestingly a very significant difference in the surface properties of the AAR could be observed in these regions (Fig. 9C) in presence of KCl. The presence of KCl also energetically favored compact AAR structure as compared to the control (0 M) (Fig. 9A). Based on the present data it can be fairly explained that KCl stabilizes the AAR structure by masking the exposed electrostatic charge on the protein. The implications can be drawn that KCl-mediated improved activity of AAR is imparted through structural stabilization as reported previously for other enzymes as well [28].
However, at higher concentration of GdnHCl acts as a classical denaturant leading to the unfolding of AAR protein chain as could be observed beyond 3.0 M concentration of GdnHCl (Fig. 8).
From the data, it could be extracted that though protein appears to resist change in the structure during unfolding till moderate concentration of denaturant interestingly the cooperativeness of the unfolding process is low as indicated by the cooperativity parameter (m) calculated by fitting the data in protein folding equation (Table 3). The low value of m indicates that the unfolding process follows multiple transitions rather than a single cooperative transition [29] which contributes to the marginal stability of the protein. This low cooperativity in unfolding could be attributed to the recently reported intrinsic flexible regions in the AAR protein [15]. In the E. coli proteome, it had been found that approximately 15% of proteins have a free energy change of ≤ 4.0 kcal/mol, while it undergoes an abrupt decrease once the cell is in thermal stress 25 [20]. Hence, the above-calculated values of thermodynamic stability parameters for AAR reflect that it is a marginally stable entity (Table 2 and 3), which can be destabilized by applied minimal stress to the cell. The study further proposes that the borderline stability of the AAR protein makes it prone to aggregation when expressed in E. coli cells for hydrocarbon production and in turn affecting the hydrocarbon production level. Hydrocarbon production levels have been reported to fluctuate with varying soluble levels of AAR [14].
This highlights the fact that the AAR protein aggregation could be one of the limitations associated with reduced hydrocarbon production, which in the present study could be correlated with the marginal stability of the protein when expressed in E. coli. Thus, the outcome of the present study involving the unfolding pathways, determination of thermodynamic stability provides us preliminary information about the stability and mechanistic aspects of the unfolding of a key nodal enzyme involved in biosynthetic hydrocarbon production pathway. Taking into consideration the present studies and further careful investigation in the near future, it would certainly provide muchfocused information on improving the stability of the enzyme in question, in turn, to further improve hydrocarbon production.

Conclusion
The present work investigated the crucial structural aspects of AAR protein associated with its stability and folding, which in-turn reflected on structural integrity of the protein. mediates structural stabilization and regional alteration towards the binding site of its neighbouring pathway enzyme aldehyde deformylating oxygenase. Based on these evidences, we propose that the marginal stability of AAR are plausible contributing reasons for aggregation propensity and hence low catalytic activity of the enzyme when expressed in E. coli for biofuel production.

Expression and purification of AAR
The M15 strain of E. coli was transformed with the pQE30AAR plasmid containing codon optimized gene encoding AAR from Synechococcus elongatus PCC7942 [8].

Data analysis of two-state equilibrium unfolding
The urea-induced equilibrium unfolding data obtained from CD and fluorescence spectroscopy was analyzed by drawing the baseline for the native and unfolded states in the unfolding transition curve which can be further fitted to a two-state model. The proposed equilibrium is Where KNU is the equilibrium constant for N ⥨ U In general SN, SI and SU are dependent on the concentration of c, and we assume a linear dependence on the concentration of denaturant (c), as SN = a1 + b1c, SI = c1 + p1c and SU = e1+ g1c where a1, b1, c1, p1, e1 and g1 are constants. a1, b1, c1, p1, e1, g1 were calculated from intercept and slope of the baseline of the native, intermediate, and unfolded states. By using all the above equations, the GdnHCl-mediated equilibrium unfolding data were analyzed by a method of nonlinear least-squares analysis and thermodynamic parameters of stability were calculated.

Molecular dynamics (MD) simulations
The crystal structure of AAR was obtained from the PDB entry 6JZU. The obtained structure was prepared before the simulation using Chimera tool [30]. The molecular dynamics on 298 K and 35 303 K were performed using AMBER18 software [31,32] and modelling and data analysis were performed by AmberTools18 suite of programs [31,32]. The parameters and atom-types of hexadecanal ligand were generated through ANTECHAMBER module and charges calculated using AM1-BCC method and the topology and parameter files were constructed using the force field leaprc.ff99SBxildn for the AAR protein.

Consent for publication
Not Applicable

Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information files.

Competing interests
The authors declare that they have no competing interests         Table 3.