Hexamerization of Geranylgeranylglyceryl Phosphate Synthase

Mar 30, 2018 - The cell membranes of all archaea contain ether lipids, and a number of archaea are hyperthermophilic. Consequently, the enzymes that ...
0 downloads 0 Views 14MB Size
Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES

Hexamerization of geranylgeranylglyceryl phosphate synthase ensures structural integrity and catalytic activity at high temperatures Mona Linde, Kristina Heyn, Rainer Merkl, Reinhard Sterner, and Patrick Babinger Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.7b01284 • Publication Date (Web): 30 Mar 2018 Downloaded from http://pubs.acs.org on March 30, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Hexamerization of geranylgeranylglyceryl phosphate synthase ensures structural integrity and catalytic activity at high temperatures Mona Linde1, Kristina Heyn1, Rainer Merkl1, Reinhard Sterner1, and Patrick Babinger1* 1

Institute of Biophysics and Physical Biochemistry, University of Regensburg, 93040

Regensburg, Germany

*Corresponding author. Phone: +49 (941) 943 1634; Fax: +49 (941) 943 2813; E-mail: [email protected]

1 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 41

ABSTRACT The cell membranes of all archaea contain ether lipids, and a number of archaea are hyperthermophilic. Consequently, the enzymes that catalyze the synthesis of membrane ether lipids had to adopt to these rough conditions. Interestingly, the enzyme that establishes the first ether bond in these lipids, the geranylgeranylglyceryl phosphate synthase (GGGPS), forms hexamers in many hyperthermophilic archaea, while also dimeric variants of this enzyme exist in other species. We used Methanothermobacter thermautotrophicus GGGPS (mtGGGPS) as a model to elucidate the benefit of hexamerization. We studied the oligomerization interfaces in detail by introducing disturbing mutations and subsequently compared the stability and activity of generated dimeric and monomeric variants with the wild type enzyme. Differential scanning calorimetry revealed a biphasic denaturation of mtGGGPS. The temperature of the first transition varies and rises with increasing oligomerization state. This first phase of denaturation leads to catalytic inactivation, but CD spectroscopy indicated only minor changes on secondary structure level. The residual part of the fold is extremely thermostable and denatures in a second phase at temperatures > 120 °C. The analysis of another distant native GGGPS enzyme affirms these observations. Molecular Dynamics simulations revealed three structural elements close to the substrate binding sites with elevated flexibility. We assume that hexamerization might stabilize these structures, and kinetic studies support this hypothesis for the binding pocket of the substrate glycerol 1-phosphate. Oligomerization might thus positively affect the thermostabilityflexibility-tradeoff in GGGPS by allowing a higher intrinsic flexibility of the individual protomers.

2 ACS Paragon Plus Environment

Page 3 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

INTRODUCTION The membrane lipids of Archaea differ significantly in their composition from those of Bacteria and Eukaryotes. They consist of polyprenyl moieties that are linked to a glycerol 1-phosphate (G1P) backbone via ether bonds, while bacterial and eukaryotic membrane lipids are built from fatty acids that are esterified to glycerol 3-phosphate (G3P). The first step in the synthesis of archaea-type ether lipids, the transfer of a geranylgeranyl residue (C20; four isoprene units) to G1P, is catalyzed by the geranylgeranyglyceryl phosphate synthase (GGGPS). Consequently, GGGPS has long been regarded to be archaea-specific and is discussed to be one of the key enzymes in the evolutionary separation of the three domains of life1-9. However, we recently showed that also some bacteria are using GGGPS-like enzymes to synthesize ether lipids with either four (Bacteroidetes) or seven (Bacillales) isoprene units, albeit with still unknown function9-12. The GGGPS enzyme family can be divided into two groups, both comprising archaeal and bacterial members (Figure 1). Group I contains the enzymes of Bacillales and of some Euryachaeota, among them all Halobacteria, while group II contains the large majority of Archaea and the Bacteroidetes. In a comprehensive study of representative members among the most important phylogenetic orders9, we analyzed their oligomerization states (as indicated by grey symbols in Figure 1) and solved some crystal structures. Additionally, we evaluated sequence similarity networks and deduced motifs from multiple sequence alignments that represent the different oligomerization states that we have found. Our findings strongly suggest that all members of group I GGGPS family enzymes exist as dimers, whereas among group II, both dimeric and hexameric complexes occur. A large number of Euryarchaeota and Crenarchaeota, among them the large majority of hyperthermophilic species, most likely possess a hexameric GGGPS9, while among species thriving at moderate temperatures, both hexameric and dimeric GGGPS occur. For many years, oligomerization has been regarded as a main factor contributing to thermostability13, 14. Here, we used the GGGPS family to study the thermostabilizing benefits of oligomerization. For this purpose, we performed a comparative analysis of wild type proteins as well as of dimeric and monomeric variants generated from hexameric Methanothermobacter

3 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 41

thermautotrophicus GGGPS (mtGGGPS) by site-directed mutagenesis. MtGGGPS shows a biphasic thermal denaturation. The temperature of the first transition is dependent on the oligomerization state and goes along with loss of catalytic activity. The residual part of the mtGGGPS fold is extremely thermostable and denatures in a second phase at temperatures >

120 °C. The results point to a stabilizing effect of oligomerization for flexible structural elements close to the active site, which we have identified by Molecular Dynamics (MD) simulations. The analysis of a GGGPS from a distant species showed that this kind of stabilization might be widespread among this enzyme family.

Figure 1: Phylogenetic tree of the GGGPS enzyme family. To calculate the tree, one representative sequence from each relevant phylogenetic order that possesses a GGGPS-like protein was selected. Oligomerization states (indicated by 2-6 grey rectangles) have been determined by static light scattering for several representative enzymes. Phylogenetic orders are given in brackets, respectively phyla, if the order is not defined. Figure adopted from Peterhoff et al.9.

4 ACS Paragon Plus Environment

Page 5 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

MATERIALS AND METHODS Cloning and site-directed mutagenesis – As summarized in Table S1, eleven mutants of Methanothermobacter

thermautotrophicus

GGGPS

(mtGGGPS)

were

generated

by

QuickChange mutagenesis15. Another mtGGGPS mutant and the wild type GGGPS genes from Methanothermobacter thermautotrophicus, Flavobacterium johnsoniae, Spirosoma linguale, Thermococcus kodakaraensis and Thermoplasma acidophilum have been cloned previously9. All genes were cloned via the NdeI/XhoI restriction sites into the expression vector pET21a (Novagen), providing a C-terminal hexahistidine (His)6 tag, and the constructs were verified by sequencing. The sequence numbering of mtGGGPS used in this study refers to EMBL ENA entry AAB85058 and pdb-id 4mm1, which have an N-terminal three amino acid extension compared to Uniprot entry O26652. Production and purification of recombinant proteins - Heterologous gene expression was performed in the E. coli strain BL21-CodonPlus(DE3)-RIPL (Agilent Technologies). The transformed cells were grown at 37 °C in LB containing ampicillin (150 µg ml-1) and chloramphenicol (30 µg ml-1). When OD600 reached 0.6–0.8, expression was induced by adding 1 mM isopropyl-β-D-1-thiogalactopyranoside (IPTG), and growth was continued overnight. After harvesting by centrifugation, cells were resuspended in either 50 mM potassium phosphate, pH 7.5, 300 mM KCl, 10 mM imidazole or 50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 10 mM imidazole and disrupted by sonication. The His-tagged proteins were purified from the clarified cell extract by metal chelate affinity chromatography. An ÄKTApurifier system with a HisTrap FF crude column (5 ml, GE Healthcare) was used, and a linear gradient of imidazole (10–500 mM) in 50 mM potassium phosphate, pH 7.5, 300 mM KCl or 50 mM Tris-HCl, pH 8.0, 300 mM NaCl was applied to elute the protein. Interfering imidazole and salt were removed from the purified proteins by dialysis against 50 mM potassium phosphate, pH 7.5 or 50 mM TrisHCl, pH 8.0 at 4 °C. The His-tag was not cleaved off after purification. Protein concentrations were determined by absorbance spectroscopy. The molar extinction coefficients ε280 and the molecular weight were calculated from the amino acid sequence by means of ProtParam16. Purified protein was dropped into liquid nitrogen and stored at -80 °C. If high protein concentrations were needed, the samples were concentrated by ultrafiltration (Amicon Ultra-15, mwco 10 kDa, Merck Millipore). The purification yields of all proteins are listed in Table S1.

5 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 41

Characterization of oligomerization state of recombinant proteins – All variants were characterized by size exclusion chromatography (SEC) experiments. Standard SEC experiments were performed on a calibrated Superdex S75 10/300 GL column (GE Healthcare), which was operated with 50 mM potassium phosphate, pH 7.5, 300 mM KCl at a flow rate of 0.5 ml min−1. 100 µl of protein with a subunit concentration of 15-200 µM (40 µM subunit concentration for standard runs) was applied. The apparent molecular weights of mtGGGPS_W141A (dimer) and mtGGGPS_A162E_W141A (monomer) as deduced from SEC calibration runs correlated well with the molecular weights as calculated from the amino acid sequence (mtGGGPS_W141A: apparent 23.8 kDa, calculated 27.5 kDa; mtGGGPS_A162E_W141A: apparent 51.3 kDa, calculated 55.0 kDa). The hexameric oligomerization state of mtGGGPS_wt has been analyzed in detail by SEC-SLS9. This allowed us to use these variants as references for their oligomerization states in subsequent SEC experiments. To identify the oligomerization state at temperatures above the first transition temperature, samples (40 µM, 200 µl) were incubated at elevated temperature (mtGGGPS_wt: 100 °C, 20 min; mtGGGPS_W141A: 70 °C, 10 min; mtGGGPS_A162E_W141A: 55 °C, 10 min), centrifuged (16,100g, 15 min) and the supernatant was analyzed on a calibrated Superdex S200 10/300 GL column (GE Healthcare) as described above. Unheated samples served as reference. Circular dichroism (CD) spectroscopy - Spectra of 10 μM protein (subunit concentration) in degassed 50 mM potassium phosphate pH 7.5 were recorded at a scan rate of 50 nm/min with a response time of 0.5 sec from 190 to 260 nm in a JASCO J-815 circular dichroism spectrometer (d = 0.1 cm). Additionally, the protein samples were heated from 25 °C to 95 °C at a ramp rate of 1 K min–1, and the CD signal was recorded at a fixed wavelength. Spectra at 25 °C were recorded before ramp heating, spectra at 95 °C were recorded immediately after ramp heating. Averaged spectra of three measurements are shown. Differential scanning calorimetry (DSC) - 20 μM protein (subunit concentration) was heated in degassed 50 mM potassium phosphate pH 7.5 from 30 °C to 130 °C at a ramp rate of 1 K min–1 in a VP-DSC differential scanning microcalorimeter (MicroCal, Malvern Intruments) with fixed reference and sample cell (0.545 ml each). The change in heat capacity with raising temperature was recorded. The proper equilibration of the calorimeter was ascertained by performing several buffer-buffer baselines. Overpressure was applied to prevent boiling above 100 °C. DSC

6 ACS Paragon Plus Environment

Page 7 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

experiments were evaluated using the Origin™ software supplied by the instrument manufacturer. Baselines were corrected using the ‘cubic connect’ option and the data was fitted with a non-two state model. The apparent midpoint temperature (TMapp) of the irreversible unfolding transition was determined as an operational measure of protein stability. The experiments were done in duplicates. Enthalpy changes of the observed first transitions were calculated as the area under the curve from single experiments. For this purpose, the heat capacity of the sample was integrated over the temperature range of the first transition peak, as defined by the zero baseline. Differential scanning fluorimetry (nanoDSF) - 20 μM protein (subunit concentration) was heated in degassed 50 mM potassium phosphate pH 7.5 from 20 °C to 95 °C at a ramp rate of 1 K min–1 in a Prometheus NT.48 instrument (NanoTemper Technologies GmbH; access provided by 2bind GmbH). The excitation power at 280 nm was 10 % and emission spectra were measured at 330 and 350 nm. The change in the ratio of the fluorescence signal at 350 nm to 330 nm with raising temperature was followed. Fluorescence transitions were fitted by the program supplied by the manufacturer. The apparent midpoint temperatures (TMapp) of the irreversible unfolding transition, defined by the inflexion points of the curves, were determined as an operational measure of protein stability. The experiments were done in duplicates. Radiometric GGGPS activity assays -

14

C-G1P was synthesized as described by Guldan et al.

(2011). To test the activity of purified GGGPS enzymes, 1 μM of protein (subunit concentration) was incubated with 20 μM GGPP and 20 μM

14

C-G1P (302 nCi) in 10 mM MgCl2, 0.2 %

Tween80, 50 mM Tris-HCl, pH 8.0, in a total volume of 100 µl for 2 h at 40 °C. The products were dephosphorylated by adding 1 U calf intestinal alkaline phosphatase (New England Biolabs) for 1 h at 40 °C and extracted according to the method of Bligh and Dyer17 as modified by Kates18. The solvent was evaporated to dryness in a rotary evaporator and the products were analyzed by TLC on Silica 60 plates, developed in ethyl acetate/hexane 1:1 (v/v), and visualized with a phosphorimager system (PerkinElmer Life Sciences). Steady-state enzyme kinetics - For determining kinetic constants, a continuous enzyme-coupled assay for phosphate detection was used19. The activity of GGGPS enzymes was assayed in the presence of GGPP and G1P. The assay mixture contained 50 mM Tris-HCl, pH 8.0, 10 mM MgCl2, 0.2 % Tween80, 1.25 mM inosine, and 0.027, 0.25, and 2.5 U/mL of Escherichia coli

7 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 41

pyrophosphatase (PPase), bacterial purine nucleoside phosphorylase (PNPase), and microbial xanthine oxidase (XOD; all enzymes obtained from Sigma-Aldrich), respectively. For GGPPdependent kinetics, GGPP (0.1-10 μM) and G1P (350 μM for mtGGGPS_wt or 3 mM for mtGGGPS_W141A) were mixed with the assay mixture in a total volume of 200 µl and incubated at 40 °C. For G1P-dependent kinetics, G1P (0.3-5000 μM) and GGPP (11 μM for mtGGGPS_wt or 7 µM for mtGGGPS_W141A) were mixed with the assay mixture in a total volume of 200 µl and incubated at 40 °C. The reaction was started by addition of enzyme (100 nM for mtGGGPS_wt or 500 nM for mtGGGPS_W141A), and monitored at 293 nm using a Jasco V650 spectrophotometer (d = 1 cm). At this wavelength, the ε of uric acid was assumed equal to 12.6 × 103 M-1cm-1. Reaction velocities were calculated from the protein concentration and the initial slopes. Kinetic constants were obtained by fitting the Michaelis-Menten equation to the data, using SigmaPlot 12.0. Average values of triplicates were calculated. Irreversible thermal inactivation - For determining kinetics of irreversible thermal inactivation, a continuous enzyme-coupled assay for phosphate detection was used19. All enzymes were adjusted to a concentration of 40 µM in 50 mM Tris-HCl, pH 8.0 (at room temperature). Actually, the pKa value of Tris buffer decreases 0.031 units per 1 K increase of temperature and therefore pH values at the denaturating temperature of highly stable proteins is up to 2.3 pH units lower (at 100 °C: pH 5.7, at 85 °C: pH 6.2, at 75 °C: pH 6.5, at 65 °C: pH 6.8 and at 50 °C: pH 7.3)20. Tris-HCl buffer had to be used because potassium phosphate buffer would disturb the phosphate detection assay. 100 µl aliquots were incubated at given temperature for 0-120 min. The samples were chilled on ice and centrifuged at 16,100 g for 5 min at 4 °C. The residual maximum velocity of the supernatant was assayed in the presence of GGPP and G1P. The assay mixture contained 50 mM Tris-HCl pH 8.0, 10 mM MgCl2, 0.2 % Tween80, 1.25 mM inosine, 300 µM G1P, 10 µM GGPP and 0.027, 0.25, and 2.5 U/mL of PPase, PNPase, and XOD, respectively, in a total volume of 200 µl and was incubated at 40 °C. The reaction was started by addition of 1 µl enzyme (200 nM final concentration, resulting from a 40 µM enzyme sample) or 6 µl enzyme (1.2 µM final concentration, for mtGGGPS_W141A), and monitored at 293 nm using a Jasco V650 spectrophotometer (d = 1 cm). At this wavelength, the ε of uric acid was assumed equal to 12.6 × 103 M-1cm-1. Reaction velocities were calculated from the initial slopes. The half-life t1/2 was obtained by fitting the exponential decay equation to the data, using SigmaPlot 12.0. Average values of triplicates were calculated.

8 ACS Paragon Plus Environment

Page 9 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

For thermal inactivation studies of the monomeric variant mtGGGPS_A162E_W141A, radiometric assays had to be used due to the low enzymatic activity of this variant. The protein was diluted to a concentration of 40 µM in 50 mM potassium phosphate, pH 7.5. 100 µl aliquots were incubated at given temperatures (45 °C, 50 °C, 55 °C, 60 °C) for 0-15 min. The samples were chilled on ice and centrifuged at 16,100 g for 5 min at 4 °C. The residual maximum velocity of the supernatant was assayed in the presence of GGPP and G1P. The assay mixture contained 20 μM GGPP and 20 μM

14

C-G1P (302 nCi) in 10 mM MgCl2, 0.2 % Tween80, 50

mM Tris-HCl pH 8.0 in a total volume of 100 µl. The reaction was started by addition of 25 µl enzyme (10 µM final concentration, resulting from a 40 µM enzyme sample) and incubated for 2 h at 40 °C. To allow separation by TLC, the products were dephosphorylated by adding 1 U calf intestinal alkaline phosphatase for 1 h at 40 °C and extracted according to the method of Bligh and Dyer17 as modified by Kates18. The solvent was evaporated to dryness in a rotary evaporator and the products were analyzed by TLC on Silica 60 plates, developed in ethyl acetate/hexane 1:1 (v/v), and visualized with a phosphorimager system (PerkinElmer Life Sciences). To determine tapp1/2 values, the intensities of the individual product spots were quantified using the Software OptiQuant 3.0 (PerkinElmer Life Sciences) and plotted against the incubation time. Molecular Dynamics (MD) simulations - MD simulations were conducted with Yasara21, version 16.9, force field AMBER03, based on one monomer deduced from the crystal structure of mtGGGPS (pdb-id 4mm1). A simulation cell was created, which was 5 Å larger than the protein along each axis and filled with water to a density of 0.997 g/ml, and counterions were added to a final concentration of 0.9 % NaCl. The simulation was conducted at three different temperatures, namely 25 °C, 45 °C, and 80 °C, adjusted by the Berendsen thermostat. For each temperature, three independent simulation runs of length 100 ns were performed with slightly different temperatures (± 0.1 °C) to alter starting conditions. Each minimization consisted of an initial equilibration step of length 14 ps followed by a heating step to reach the end temperature within 4 ps. During the next 100 ns of simulation, a snapshot was recorded every 20 ps. The snapshot-specific distribution of secondary structure elements of the full protein was determined by means of the function Yasara:SecStrAll. Corresponding values originating from the three runs conducted in parallel for each temperature were averaged. The root mean square fluctuation (RMSF) value representing for each C-atom the deviation from the mean position

9 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 41

was deduced for each series of snapshots. For this calculation, the function R:rmsf was used; corresponding values resulting from the three temperature-specific runs were averaged. By means of PyMol22, RMSF values were mapped onto the 3D structure and visualized. RMSF values were read as B-factors and a color gradient ranging from blue over yellow to red was used to indicate values in the range [0 Å, ≥3 Å]. For statistical analyses, a reference RMSF value was determined for each C-atom i as the C

C

mean ( RMSFT25 ,i , RMSFT45 ,i ) resulting from the RMSF values determined at temperatures T =

25 °C and 45 °C. To characterize the deviation dev(C ,i ) of the RMSF values at 80 °C, a z-score was calculated according to dev(C ,i ) 

norm(C ,i )  norm(C ,i )



.

norm ( C ,i )

Here, norm (C ,i )  RMSFTC80 ,i  mean ( RMSFTC25 ,i , RMSFTC45 ,i ) is the difference of the RMSF values at 80 °C and the reference RMSF values. For the calculation of the standard deviation, the N-terminal residues 1 - 12 and the C-terminal residues 226 - 245 were discarded.

10 ACS Paragon Plus Environment

Page 11 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

RESULTS We chose the GGGPS from Methanothermobacter thermautotrophicus (mtGGGPS) as a model protein to study the impact of hexamerization within the GGGPS family, because its crystal structure has been solved (pdb-id 4mm1), it shows high activity and it is extremely thermostable.9 GGGPS enzymes exhibit the (βα)8-barrel fold, which is frequently found in nature23, but with some characteristic variations9,

24

(Figure S1). They have an additional N-

terminal helix α0, helix α3 is reduced (α3*) or even completely replaced by a strand, and helix α5 is split. Helix α5’ is lying on top of the barrel, where the active site is located. The hexameric mtGGGPS protein consists of six identical protomers (Figure 2). Each two of them form dimer modules that possess the same topology like native dimeric GGGPS proteins9. Three of those dimer modules associate to the hexamer. Therefore, the hexamer can be regarded as a trimer of dimers. The subunits assemble in a way that two rings (an “upper” and a “lower” ring) of three protomers each are formed (Figure 2A, bottom line; Figure 2B). Three different interface types connect the subunits: (i) The dimer module interface (Figure 2B, colored dotted lines) is symmetric, is identical to the dimerization interface of native dimeric GGGPS and connects two protomers each. (ii) The interconnecting interface (black curved lines) is also symmetric and connects protomers of two different dimer modules each. Together with the dimer module interface, it forms a zig-zag-like connection between all six protomers. Both connect the upper and the lower ring with each other. (iii) The ring interface is asymmetric (magenta lines with arrows). Three such interfaces connect one subunit of each of the dimeric modules to form the upper and the lower ring, which are staggered on gap (60°) against each other. The dimer module interface is the largest one (~ 1160 Å2), followed by the interconnecting interface (~ 620 Å2) and the quite small ring interface (~ 260 Å2). To identify relevant residues within the contact interfaces and to study the benefit of hexamerization, we introduced mutations to disturb the interactions and subsequently analyzed the mtGGGPS variants with respect to their oligomerization state, stability and activity. The mutant genes were recombinantly expressed in E. coli, and the purity of all proteins was estimated to be > 95 % by SDS-PAGE (Figure S2). Furthermore, the structural integrity was verified by CD spectroscopy (Figure S3). In the following paragraphs, we describe the properties of each interface in detail and discuss their interplay to stabilize the hexamer.

11 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 41

Figure 2: Configuration of the mtGGGPS hexamer. (A) Two mtGGGPS protomers form a dimer module, and three of those dimer modules associate to the hexamer. In the bottom line, the dimer modules are shown in different colors. One of the protomers of each dimer module is shown in pale color, where necessary for clarity. (B) Schematic and exploded view of the hexamer. The protomers are shown in different colors for each dimer module, according to (A). The three different interface types within the hexamer are represented by different symbols, as indicated in the legend.

The ring interface is highly conserved and essential for hexamerization The core of the small, asymmetric ring interface is formed by the residues W141 and D144 in helix α5’ at one interacting surface, and by K146 plus a hydrophobic cleft which is formed by small hydrophobic amino acids, mainly I155, at the opposite surface in the loop region α5’α5 (Figure 3). In a previous study9, we have postulated that the interface is stabilized by a cation-π interaction between W141 and K146, and we have shown that the mutation W141A disrupts the hexamer to dimers (dimer modules). In fact, the aromatic residue at the position of W141 is

12 ACS Paragon Plus Environment

Page 13 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

strictly conserved in all putatively hexameric GGGPS, as deduced from a multiple sequence alignment9, and a positive charge (R, K) at the position of K146 is frequent (63% conservation).

Figure 3. The asymmetric ring interface of mtGGGPS. To distinguish between the associated subunits of each dimer module, each upper subunit is shown in color, the lower in pale (left side). Important interface residues are depicted as sticks in magenta. The contact between the interacting surfaces of two adjacent protomers (blue and cyan) is shown in detail (right side).

To prove the importance of K146, we have now produced the variant mtGGGPS_K146A and analyzed its oligomerization state using analytical size exclusion chromatography (SEC) (Figure 4). As expected for a cation-π interaction, the mutation K146A also leads to dimeric proteins, like the mutation W141A does. Because K146 could also form an ionic interaction with D144, we separately mutated D144 to alanine. This mutation did not lead to a collapse of the hexameric conformation (Figure 4). Although a negative charge at the position of D144 is frequent (54% conservation), the putative ionic interaction is obviously less relevant for hexamer stabilization. Because W141 extends into a hydrophobic pocket of the opposite protomer, which is mainly formed by I155, the latter amino acid was also mutated to elucidate the importance of the hydrophobic pocket for hexamerization. Interestingly, the mtGGGPS_I155A variant obviously exists in a relatively fast association-dissociation equilibrium between the dimeric and hexameric state under the given experimental conditions (Figure 4). Analytical SEC of mtGGGPS_I155A with varying concentrations approved that this association-dissociation equilibrium is concentration-dependent (Figure S4). This shows that the interaction of the aromatic residue with the hydrophobic groove also contributes to hexamer stabilization, and it might be the most important interaction in hexamers that are not linked by a cation-π interaction, as the positive charge is less conserved. It should be noted that we reported about the residue K152, which is

13 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 41

close to the ring interface and strictly conserved in hexameric GGGPS.9 However, this residue is not involved in hexamerization (data not shown). Because K152 is also located close to the active site of an opposite protomer, we rather assume an effect on catalysis that still needs to be elucidated.

Figure 4. Analytical size exclusion chromatography of mtGGGPS variants with amino acid exchanges at the ring interface. The proteins (40 µM, subunit concentration) were applied to a S75 analytical column equilibrated with 50 mM KP pH 7.5, 300 mM KCl. Elution was performed at a flow rate of 0.5 ml/min, followed by measuring the absorption at 280 nm (A280) and plotted against the elution volume. mtGGGPS_wt (wt, dashed lines, magenta) and mtGGGPS_W141A (W141A, dashed lines, blue) are shown as references for the hexameric and dimeric oligomerization states. mtGGGPS_K146A (K146A, blue), mtGGGPS_D144A (D144A, magenta) and mtGGGPS_I155A (I155A, cyan) are drawn as solid lines. The oligomerization state is indicated by grey symbols.

Interplay between dimer module interface and ring interface The dimer module interface has previously been studied in detail on the example of the dimeric GGGPS-like protein PcrB25. In that study, a number of amino acid replacements have disrupted the dimeric complex yielding stable and functional monomers. To introduce analogous replacements in mtGGGPS, its structure (pdb-id 4mm1) was superimposed with that of Bacillus subtilis PcrB (bsPcrB; pdb-id 1viz). The positions mtGGGPS_A162 (corresponding to bsPcrB_V148) and mtGGGPS_M168 (corresponding to bsPcrB_L152) were selected for further analysis (Figure 5). A162 is localized in the center of the interface (in the middle of helix α4) and

14 ACS Paragon Plus Environment

Page 15 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

M168 at the lower part of the interface at the loop region α5β6. The two amino acids were mutated independently to glutamate in the hexameric mtGGGPS_wt. Glutamate was chosen to introduce electrostatic repulsion between the two adjacent protomers of one dimer module in order to disrupt the interface. The same was done in the background of the dimeric mtGGGPS_W141A variant to verify the monomerizing effect of the mutations. The oligomerization state of the variants was analyzed by analytical SEC. The mutation M168E did not have any effect at all and was therefore not further characterized (data not shown).

Figure 5. The symmetric dimer module interface of mtGGGPS. For clarity, only one dimeric module of the hexameric mtGGGPS is shown in survey (left side) and in detail (right side), and one of the protomers is shown in pale. By superimposition with the structure of bsPcrB, relevant interface residues were identified and are depicted as sticks in magenta.

The

incorporation

of

the

mutation

A162E

into

the

dimer

(resulting

in

mtGGGPS_A162E_W141A) leads to a monomeric protein (Figure 6). The incorporation of this monomerizing mutation into the hexameric mtGGGPS_wt (resulting in mtGGGPS_A162E), however, did not lead to a collapse of the hexameric structure. This strongly indicates that even if the dimer module interface is disturbed, the interconnecting interface is strong enough to fix together the two trimer rings of the hexamer (Figure 2). Otherwise, trimers would be expected in SEC experiments with mtGGGPS_A162E, or even monomers, assumed that the trimeric rings are not stable alone.

15 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 41

Figure 6. Analytical size exclusion chromatography of mtGGGPS variants with the A162E exchange at the core of the dimer module interface. The proteins (40 µM, subunit concentration) were applied to a S75 analytical column equilibrated with 50 mM KP pH 7.5, 300 mM KCl. Elution was performed at a flow rate of 0.5 ml/min, followed by measuring the absorption at 280 nm (A280) and plotted against the elution volume. mtGGGPS_wt (wt, dashed lines, magenta) and mtGGGPS_W141A (W141A, dashed lines, blue) are shown as reference peaks for hexameric and dimeric oligomerization state. mtGGGPS_A162E (A162E, light pink) and mtGGGPS_A162E_W141A (A162E_W141A, green) are drawn as solid lines. The oligomerization state is indicated by grey symbols.

Interplay of the interconnecting interface with the ring interface and dimer module interface The interconnecting interface always connects one subunit from the upper ring with one subunit from the lower ring, which belong to different dimer modules (Figure 2). In combination with the dimer module interface, a kind of zig-zag link is generated. As shown above, even if the dimer module interface is disturbed, the interconnecting interface is able to stabilize the hexameric conformation, signalizing that this interface is more important for hexamerization than the dimer module interface. To confirm this hypothesis, mutational analysis was performed at the interconnecting interface, where four relevant residues (D57, N82, R88, Y105) have been identified9 (Figure 7). N82 is localized at the beginning of helix α3* and forms H-bonds with Y105 of the adjacent protomer, while Y105 is localized at the beginning of helix α4 and interacts by stacking with Y105 of the adjacent protomer. R88 is localized at the end of helix α3*, and D57 in the loop βα2. D57 and R88 build polar contacts with each other.

16 ACS Paragon Plus Environment

Page 17 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 7. The interconnecting interface of mtGGGPS. To distinguish between the associated subunits of each dimer module, each upper subunit is shown in color, the lower in pale (left side). The interface is shown in survey (left side) and in detail (right side). Important interface residues are depicted as sticks in magenta.

As Y105 is participating in two interactions, this residue was subjected to mutagenic analysis. To analyze its cross-connecting properties, Y105 was mutated to the small amino acid alanine in the backgrounds of mtGGGPS_wt (resulting in mtGGGPS_Y105A) and mtGGGPS_A162E (resulting in mtGGGPS_A162E_Y105A). Additionally, the residues D57 and R88 were subjected to mutagenesis as well. As these two residues are involved in a polar contact between each other, D57 was mutated to arginine or R88 to glutamate to introduce repulsion. As for the mutation Y105A, both mutations were separately introduced into mtGGGPS_wt (resulting in mtGGGPS_D57R

and

mtGGGPS_R88E)

and

into

mtGGGPS_A162E

(resulting

in

mtGGGPS_A162E_D57R and mtGGGPS_A162E_R88E). The variants were analyzed via analytical SEC (Figure 8, Figure S5).

Figure 8. Analytical size exclusion chromatography of the mtGGGPS variants with the Y105A exchange at the interconnecting interface. The proteins were applied to a S75 analytical column equilibrated with 50 mM KP pH 7.5, 300 mM KCl. Elution was performed at a flow rate of

17 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 41

0.5 ml/min, followed by measuring the absorption at 280 nm (A280) and plotted against the elution volume. mtGGGPS_wt (wt, dashed lines, magenta, 40 µM, subunit concentration), mtGGGPS_W141A (W141A, dashed lines, blue, 40 µM, subunit concentration) and mtGGGPS_A162E_W141A (A162E_W141A, dashed lines, green, 33 µM) are shown as reference peaks for hexameric, dimeric and monomeric oligomerization state. mtGGGPS_Y105A (Y105A, blue, 25 µM, subunit concentration) and mtGGGPS_A162E_Y105A (A162E_Y105A, green, 40 µM, subunit concentration) are drawn as solid lines. The oligomerization state is indicated by grey symbols.

Incorporation of Y105A into the mtGGGPS_A162E variant resulted in monomeric proteins, whereas this mutation in mtGGGPS_wt lead to a collapse of the hexamer into dimer modules. This shows that if either the interconnecting interface or the ring interface is disturbed, the hexameric structure collapses into its dimer modules. The ring interface cannot stabilize the hexamer over the ring system, if the interconnecting interface is destroyed, and the interconnecting interface cannot crossconnect the dimer modules and stabilize the hexamer, if the ring interface is loosened. The dimer modules resulting after disturbance of one of these interfaces are of the same configuration, as they can both be dissected into monomers with the monomerizing mutation A162E. In summary, the interconnecting interface is as important as the ring interface for hexamerization. Incorporation of D57R and R88E into the hexameric mtGGGPS_wt did not lead to dimerization, but to only partial monomerization in the mtGGGPS_A162E background (Figure S5). Obviously, the deletion of the polar contact does affect the oligomerization state and is disturbing the interconnecting interface, but not as severely as the mutation of Y105. While Y105 is located near the center of the hexameric overall structure, R88 and D57 are localized further apart. One could assume that perturbation at the inner core of the hexamer does have a stronger effect on the oligomerization than disturbing at the outer shell. In the light of the high relevance of the interconnecting interface for hexamerization in mtGGGPS, it is interesting that Y105 is conserved in three of the five subgroups of hexameric GGGPS (as defined by sequence similarity network analysis9), but in the other two not. The polar contact between D57R and R88E is even less conserved. This indicates that the interconnecting interface might either be less important in other hexameric GGGPS, or that other residues are responsible for this interaction in these GGGPS variants.

18 ACS Paragon Plus Environment

Page 19 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Table 1 lists the oligomerization states of all characterized mtGGGPS variants. Table 1. Oligomerization states of mtGGGPS_wt and its variants Variant

Oligomerization state

mtGGGPS_wt

Hexamer

mtGGGPS_D57R

Hexamer

mtGGGPS_R88E

Hexamer

mtGGGPS_A162E

Hexamer

mtGGGPS_I155Aa mtGGGPS_Y105A

Hexamer – Dimerb Dimer

mtGGGPS_W141A

Dimer

mtGGGPS_K146A mtGGGPS_A162E_D57R

Dimer Hexamer – Monomerb

mtGGGPS_A162E_R88E

Hexamer – Monomerb

mtGGGPS_A162E_Y105A

Monomer

mtGGGPS_A162E_W141A

Monomer

a

The thermal stability of this variant was not analyzed b For variants that did not show a distinct oligomerization state (e.g. due to an associationdissociation equilibrium between two states), both observed states are given.

Influence of the oligomerization state on the catalytic parameters of mtGGGPS We also elucidated how oligomerization affects the catalytic parameters of the mtGGGPS variants. We recorded saturation curves for the hexameric variant mtGGGPS_wt and the dimeric variant mtGGGPS_W141A by using a colorimetric assay that detects the pyrophosphate liberated

by

the

GGGPS

reaction

(Figure

S12).

The

monomeric

variant

mtGGGPS_A162E_W141A was not accessible to such experiments due to its low activity. The deduced catalytic parameters (Table 2) of the hexameric variant mtGGGPS_wt are in good agreement with earlier results26. Interestingly, the KM (GGPP) of the dimeric variant mtGGGPS_W141A remains almost unchanged and the kcat is only slightly decreased when compared to the wild type, while the KM (G1P) is increased by a factor of ~ 50. This indicates that one function of hexamerization might be the stabilization of the G1P binding pocket by interactions between the subunits.

19 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 41

Table 2. Steady-state kinetic parameters of mtGGGPS_wt and mtGGGPS_W141Aa Variant

KM (G1P) c

[µM]

KM

kcat [s-1]

kcat/KM (G1P) -1 -1

(GGPP)

kcat/KM (GGPP)

[M s ]

[M-1s-1]

2.52 (± 3.00) · 104

6.67 (± 6.00) · 105

5.33 (± 1.54) · 104

2.38 (± 1.00) · 105

139.63 ± 54.04

2.36 (± 0.71) · 104

[µM]d mtGGGPS_wtb

mtGGGPS_wt

mtGGGPS_W141A

13.5 ± 1.0

0.51 ±

0.34 ±

0.05

0.03

1.68 ±

0.40 ±

0.10

0.01

573.0 ±

3.39 ±

0.08 ±

74.0

0.56

0.004

7.5 ± 0.7

a

The parameters were determined at 40 °C with a colorimetric assay for phosphate detection in triplicates (given with standard deviation), and fitting the Michaelis-Menten equation to the data (Figure S12). b Kinetic parameters as obtained by Soderberg et al.26 c G1P-dependent kinetics were recorded in presence of 11 μM GGPP for mtGGGPS_wt and 7 µM GGPP for mtGGGPS_W141A. d GGPP-dependent kinetics were recorded in presence of 350 μM G1P for mtGGGPS_wt and 3 mM GGPP for mtGGGPS_W141A.

Thermal stability of the mtGGGPS variants In order to compare the thermal stabilities of hexameric, dimeric and monomeric mtGGGPS variants, differential scanning calorimetry (DSC) and nano differential scanning fluorimetry (nanoDSF) were performed. DSC allows for the investigation of thermal unfolding of proteins, by measuring changes in heat capacity, to maximum temperatures of 130 °C. DSC denaturation curves of the representative variants mtGGGPS_wt (hexamer), mtGGGPS_W141A (dimer) and mtGGGPS_A162E_W141A (monomer) are shown in Figure 9, the thermograms for all other variants in Figure S6, and all data are summarized in Table 3.

20 ACS Paragon Plus Environment

Page 21 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 9. Thermal stability of mtGGGPS_wt and its variants, followed by DSC. Changes in heat capacity of 20 µM protein (subunit concentration) in 50 mM potassium phosphate, pH 7.5 were monitored from 30 °C to 130 °C at a scan rate of 1 K/min for mtGGGPS_wt (wt, magenta), mtGGGPS_W141A (W141A, blue) and mtGGGPS_A162E_W141A (A162E_W141A, green). The oligomerization state is indicated by grey symbols. Repetitive thermal denaturation experiments revealed that the thermal transitions observed by DSC are mostly irreversible (data not shown).

MtGGGPS_wt and its variants showed a characteristic DSC profile with two main peaks each, indicating two thermal transitions during unfolding. For all variants, the second transition (T2) occurs at temperatures around 120 °C. In contrast, the temperature of the first transition (T1) varies and correlates with the oligomerization state. The monomeric mtGGGPS_A162E_W141A shows a defined first transition at T1=53 °C (Figure 9), and similarly do all other monomeric variants (Figure S6A), whereas the dimeric mtGGGPS_W141A shows a transition at T1 = 68 °C (Figure 9), and similarly do all other dimeric variants (Figure S6B). The first transitions of the hexameric variants are somewhat more diverse and the peaks are broader, but they are all at significantly higher temperatures than for the dimeric variants. MtGGGPS_wt shows a double peak at 97 °C and 102 °C (Figure 9), and similarly does mtGGGPS_D57R (Figure S6C). MtGGGPS_R88E and mtGGGPS_A162E also show double peaks at somewhat lower temperatures. The reason for the double peaks remains unresolved experimentally. In case of mtGGGPS_A162E, it may be ascribed to the fact that this variant did not show a homogenous behavior in SEC runs, but might contain a fraction of monomeric proteins (Figure 6). Summarized, the DSC data indicate a uniform second transition T2 at hyperthermophilic temperatures for all variants, that we later attribute to the complete denaturation of the protein. In

21 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 41

contrast, the temperature of the first transition T1 increases with higher oligomerization state, and in the following we attribute this to the collapse of a small part of the mtGGGPS structure. To further confirm the DSC data, nano differential scanning fluorimetry (nanoDSF) was applied. This technique monitors unfolding during heating to a maximum temperature of 95 °C in a labelfree analysis, based on intrinsic tryptophan fluorescence. The nanoDSF thermograms of the representative

variants

mtGGGPS_wt

(hexamer),

mtGGGPS_W141A

(dimer)

and

mtGGGPS_A162E_W141A (monomer) are shown in Figure 10, the thermograms for all other variants in Figure S7, and all data are summarized in Table 3.

Figure 10. Thermal stability of mtGGGPS_wt and its variants, followed by nanoDSF. The change in the ratio of the fluorescence emission at 350 nm and 330 nm of 30 µM protein (subunit concentration) in 50 mM potassium phosphate, pH 7.5 was monitored from 20 °C to 95 °C at a scan rate of 1 K/min for mtGGGPS_wt (hexamer, magenta), mtGGGPS_W141A (dimer, blue) and mtGGGPS_A162E_W141A (monomer, green). The transition temperatures of mtGGGPS_W141A and mtGGGPS_A162E_W141A are marked by vertical dashed lines, the oligomerization state is indicated by grey symbols.

The monomeric variants show only one transition in the nanoDSF runs, which occurs at about 50-55 °C (Figure 10, Figure S7A), and the transition of the dimeric variants is at about 65 °C (Figure 10, Figure S7B). These temperatures are in agreement with the first transitions (T1) that have been observed in DSC experiments (Table 3). For hexameric mtGGGPS variants, no or only the onset of a transition could be observed at the upper temperature limit of 95 °C (Figure 10, Figure S7C), because according to the DSC results, the first transition only occurs at

22 ACS Paragon Plus Environment

Page 23 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

~ 100 °C in these proteins. Similarly, the second transition (T2) as observable in DSC experiments cannot be followed by nanoDSF, because it occurs at >100 °C. Interestingly, all monomeric and dimeric variants showed a decrease in the F350/F330-ratio upon the first transition, going along with a blue-shift of the fluorescence emission of tryptophan. This indicates that the tryptophans get into a more hydrophobic environment after thermal heating. This observation supports our hypothesis that the first transition (T1) in monomeric and in dimeric variants corresponds to a partial collapse of the proteins that bury the tryptophans. In contrast, upon complete denaturation buried tryptophans usually would become exposed to a more hydrophilic environment, resulting in an increase of the F350/F330-ratio upon heating in nanoDSF experiments. Only the hexameric proteins showed such an increase in the F350/F330-ratio upon thermal heating (Figure 10, Figure S7C). These results indicate that in the hexameric mtGGGPS variants, the first denaturation step may be based on other mechanisms than that in the monomeric or dimeric variants. Again as in the DSC thermograms, the mtGGGPS_A162E clearly showed two independent transitions in the nanoDSF thermogram (Figure S7C, numbered 1), with the first transition (T1) visible as decrease in the F350/F330 ratio, and the second transition (T1*) visible as an increase. As discussed above, the protein preparation of this variant obviously contains a small fraction of monomeric proteins next to the main fraction of hexameric proteins, and the two observable transitions may reflect those two fractions.

23 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 41

Table 3. Overview of the data of DSC and nanoDSF measurementsa

Variant

nanoDSFb T1 [°C] T1* [°C]

mtGGGPS_wt

> 95

mtGGGPS_D57R

> 95

mtGGGPS_R88E

> 95

mtGGGPS_A162E

53.8 ± 0.4 73.6 ± 0.1

DSCb T1 [°C] T1* [°C] 97.5 ± 0.7 102.5 ± 0.6 97.2 ± 0.1 101.7 ± 0.5 65.7 ± 1.1 89.1 ± 1.3 55.0 ± 0.0 73.0 ± 0.7

DSCb T2 [°C]

DSCc 1 transition ΔH [kJ/mol]

127.0 ± 0.5

1626

129.9 ± 0.1

781

122.2 ± 0.7

1410

125.8 ± 1.8

723

mtGGGPS_Y105A

63.2 ± 0.5

64.2 ± 1.1

124.3 ± 0.7

435

mtGGGPS_W141A

63.8 ± 0.1

68.3 ± 0.5

124.4 ± 0.1

442

mtGGGPS_K146A

66.0 ± 0.1

67.4 ± 0.2

123.5 ± 0.7

420

mtGGGPS_A162E_D57R

52.2 ± 0.1

53.8 ± 0.1

126.2 ± 0.8

368

mtGGGPS_A162E_R88E

53.9 ± 0.1

55.8 ± 0.2

114.6 ± 0.6

337

mtGGGPS_A162E_Y105A

53.4 ± 0.1

55.5 ± 0.5

122.7 ± 0.1

452

mtGGGPS_A162E_W141A

51.0 ± 0.1

52.7 ± 0.1

120.7 ± 1.0

327

st

a

Hexameric oligomerization state is depicted in magenta, dimeric in blue and monomeric in green. b T1/T1*/T2: apparent midpoint temperature, thermal transitions with T1 referring to the first observable transition, T2 referring to the last observable transition (complete unfolding of the protein) and T1* referring to observable transitions in between, average values of duplicates were calculated (given with standard deviations) c The enthalpy change of the first transition of unfolding was calculated as the area under the first transition peak from the baseline-corrected DSC data shown in Figures 9 and S6 and is given per mol of subunit.

Additionally, we analyzed the thermal denaturation of mtGGGPS variants by circular dichroism (CD) spectroscopy before and after heating to 95 °C (Figure 11). In accordance with the DSC and nanoDSF data (T1 = 97-102 °C), mtGGGPS_wt shows almost no change of the CD spectrum after heating to 95 °C, if compared to spectra obtained at 25 °C (Figure 11A). The CD spectra of all other variants show a minor, but characteristic shift upon heating to 95 °C (Figure 11B-D). All variants still give a spectrum characteristic of a folded protein, yet with a decreased α-helical,

24 ACS Paragon Plus Environment

Page 25 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

but increased random coil content. This is consistent with our hypothesis that the first transitions observed in DSC and nanoDSF experiments refer to a partial collapse, but not to a complete denaturation of the proteins. A temperature-resolved series of CD spectra on the example of mtGGGPS_W141A shows that the change in the spectrum only sets on when temperatures rise above ~ 62 °C (Figure 11E). This is at a slightly lower temperature than what has been observed as T1 in DSC and nanoDSF analysis (Table 3), but it must be noted that for recording CD spectra the protein is kept for much longer time at elevated temperatures.

25 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 41

Figure 11. Thermal denaturation of mtGGGPS variants observed by CD spectroscopy. (A-D) Far-UV CD spectra of 10 µM protein (subunit concentration) were recorded in 50 mM potassium phosphate, pH 7.5 from 185 nm to 260 nm (d = 1 mm) at room temperature (solid lines) and immediately after heating to 95 °C (dashed lines). Each spectrum represents an individual mtGGGPS variant, but labeling was omitted for clarity. The oligomerization state is indicated by

26 ACS Paragon Plus Environment

Page 27 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

grey symbols. (A) mtGGGPS_wt at room temperature (solid magenta line) and immediately after heating to 95 °C (dashed light pink line). (B) All monomeric mtGGGPS variants (A162E_D57R, A162E_R88E, A162E_Y105A, A162E_W141A) at room temperature (solid green lines) and immediately after heating to 95 °C (dashed pale green lines). (C) All dimeric mtGGGPS variants (Y105A, W141A, K146A) at room temperature (solid blue lines) and immediately after heating to 95 °C (dashed light blue lines). (D) All hexameric mtGGGPS variants (D57R, R88E, A162E) at room temperature (solid magenta lines) and immediately after heating to 95 °C (dashed light pink lines). (E) Far-UV CD spectra of the mtGGGPS_W141A variants were recorded as above at the indicated temperatures.

Repetitive thermal denaturation experiments revealed that the first transitions observed by DSC and CD spectroscopy are mostly irreversible. However, the protein solutions remained clear after heating above the temperatures of the first transitions, and UV/Vis spectra remained unchanged. To shed light on the irreversible nature of the first denaturation step, we analyzed samples before and after heating above the temperatures of the first transitions by size exclusion chromatography and found the formation of soluble aggregates that elute within the exclusion volume (Figure S8). Furthermore it must be noted that due to the irreversibility of the first transition, the DSC and nanoDSF experiments could not be performed under equilibrium conditions. Consequently, the observed transition temperatures (T1) depend on the slope of the temperature ramp during heating and must be regarded as apparent values. Tests with a subset of five variants revealed that the apparent transition temperature increases uniformly for all variants by 1.4 K (± 0.2 K) when the ramp is set from 1K/min to 2 K/min, and decreases uniformly by about 1.3 K (± 0.3 K) when the ramp is set to 0.5 K/min. Therefore, apparent transition temperatures obtained at identical temperature ramps allow for a comparison between individual variants. We also calculated enthalpy changes (ΔH) for the first transitions (T1) observed in the DSC experiments and found that they increase with the oligomerization state (Table 3). However, these data have to be taken with care: As mentioned above, partial aggregation occurs in the course of the first transition, and the enthalpy term that is contributed by aggregation might be different for the individual variants. To sum up the results from thermal denaturation experiments, DSC revealed a biphasic unfolding of all mtGGGPS variants. The temperatures of the first transition (T1) could be well reproduced by nanoDSF experiments. However, CD spectra revealed only a small signal shift after heating to 95 °C, indicating that the first transition is presumably due to the loss of only some α-helical

27 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 41

parts and leads to soluble aggregates. For technical reasons, an independent second transition (T2) at > 120 °C was observed only in the DSC experiments, which most likely represents the complete denaturation of the proteins.

Correlation of thermal stability and thermal inactivation We next analyzed whether the transitions observed in DSC and nanoDSF experiments correlate with a loss of catalytic activity. A qualitative analysis revealed that all mtGGGPS variants listed in Table 3 were moderately to highly active (Figure S9). For further analysis, mtGGGPS_wt (hexamer), mtGGGPS_W141A (dimer) and mtGGGPS_A162E_W141A (monomer) were again selected as representative variants for each observed oligomerization state. To investigate their stability against irreversible inactivation by heat, apparent half-live times (tapp1/2) were obtained by incubating the enzymes at various elevated temperatures and determining the residual activity after different time intervals (Table 4). For mtGGGPS_wt and mtGGGPS_W141A, which both showed high activity, inactivation kinetics were recorded by means of a spectroscopic assay. Time courses of the heat-induced inactivation were monitored, and tapp1/2 values were deduced from the exponential decays (Figure S10). Due to its low activity, the heat inactivation of mtGGGPS_A162E_W141A was followed using a radiometric assay, and tapp1/2 values were estimated from a densitometric analysis of the autoradiograms (Figure S11). We aimed to identify temperatures where the three variants showed comparable tapp1/2 values, and were successful in finding experimental conditions with tapp1/2 ≈ 15 min. While we had to destabilize mtGGGPS_wt by adding 0.1 M GdmCl to obtain tapp1/2 = 16.5 min at 100 °C, mtGGGPS_W141A showed a tapp1/2 = 15 min at 65 °C, and for mtGGGPS_A162E_W141A we estimate a tapp1/2 = 15 min slightly above 45 °C. Interestingly, these temperatures correlate well with the temperatures of the first transitions (T1, T1*) observed in the DSC and nanoDSF experiments (Table 4). This result is in line with our hypothesis: The first transitions of thermal unfolding lead to catalytic inactivation, but go along with only a slight change in CD spectra and represent the denaturation of only some α-helical parts of the protein. These parts are stabilized by oligomerization, resulting in increased first transition temperatures. The rest of the mtGGGPS fold is extremely thermostable, which becomes visible as a second transition in DSC experiments at very high temperatures.

28 ACS Paragon Plus Environment

Page 29 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Table 4. Correlation of heat inactivation and thermal denaturation studies of mtGGGPS_wt, mtGGGPS_W141A and mtGGGPS_A162E_W141Aa Variant

Oligomerization state

tapp1/2 [°C]:[min]

DSC T1 [°C] T1* [°C]

nanoDSF T1 [°C]

mtGGGPS_wt

hexamer

100°C: 16.5*,b

97.5 ± 0.7 102.5 ± 0.6

> 95

mtGGGPS_W141A

dimer

65°C: 15.0b

68.3 ± 0.5

63.8 ± 0.1

mtGGGPS_ A162E_W141A

monomer

45°C: > 15 50°C: 2 55°C: < 1c

52.7 ± 0.1

51.0 ± 0.1

a

DSC and nanoDSF values are from Table 3. t 1/2 was obtained by fitting a single exponential decay equation to the data from kinetic thermal inactivation studies (Figure S10). c Values estimated from a densitometric analysis of a radiometric assay (Figure S11). The data indicate that a tapp1/2 = 15 min is reached slightly above 45 °C. * 0.1 mM GdmCl was added before heat incubation b app

29 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 41

Thermal stability of other native hexameric and dimeric GGGPS To elucidate whether the observed denaturation characteristics also occur in other native GGGPS enzymes, we characterized the hexameric proteins tkGGGPS and slGGGPS as well as the dimeric proteins taGGGPS and fjGGGPS by means of DSC (Table 5; the table includes the species names; Figure S13A). Additionally, thermal inactivation experiments were conducted and CD spectra were recorded as described above. Hexameric tkGGGPS turned out to be at least as thermostable as mtGGGPS (Figure S13A,B); its tapp1/2 is very high (>>120 min) even at 100 °C. No clearly separated T1 and T2 transitions are observable in DSC, but only one transition at 115 °C with a shoulder at 111 °C, indicating that the two transitions conincide in this variant. Dimeric taGGGPS shows only a single transition at 76 °C that leads to complete denaturation, as verified by CD spectroscopy (Figure S13A,C). A dimeric state is obviously sufficient to provide the required stability in moderately thermophilic organisms like Thermoplasma acidophilum. Hexameric slGGGPS behaves similar to taGGGPS, although Spirosoma linguale is a mesophilic species. It also shows only a single transition at 83 °C that leads to complete denaturation (Figure S13A,C). Obviously, the extremely high thermostability of the fold has been lost in these species. In contrast to these proteins, dimeric fjGGGPS behaves like the hyperthermostable mtGGGPS. It shows two transitions in the DSC analysis and only a slight change of the CD signal in the course of the first transition (T1) (Figure S13D). The first transition (T1) could be verified by nanoDSF (Figure S13E). The scaffold of the protomer is again highly thermostable (T2 = 110 °C), whereas the first transition (T1 = 54 °C) correlates with the loss of activity in thermal inactivation experiments. Interestingly, fjGGGPS is a dimer and is quite distant from mtGGGPS in phylogenetic analysis (Figure 1), with only 32 % sequence identity (52 % sequence similarity). It shows the lowest T1 value observed for all GGGPS variants within this study, and Flavobacterium johnsoniae has its growth optimum at only 20 °C. Summarized, these results make clear that the occurrence of two transitions during thermal denaturation and the existence of a highly thermostable protein scaffold is even retained among some dimeric GGGPS and among some mesophilic species. However, it is not a common feature of all GGGPS, as other species have abandoned the extremely high thermostability of the fold (slGGGPS, taGGGPS).

30 ACS Paragon Plus Environment

Page 31 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Table 5. Characteristics of native hexameric and dimeric GGGPS variants

Variant

Domain Species

Oligomerization state

Growth optimum [°C]

DSCb T1 [°C]

DSCb T2 [°C]

tapp1/2 [min]

DSCe 1st transition ΔH [kJ/mol]

mtGGGPSa

Archaea M. thermautotrophicus

Hexamer

75

97.5 ± 0.7

127.0 ± 0.5

100°C: 16.5*,c

1626

tkGGGPS

Archaea Thermococcus kodakaraensis

Hexamer

85

n.o.

115.0 ± 0.0

100°C: >>120d

n.o.

taGGGPS

Archaea Thermoplasma acidophilum

Dimer

59

n.o.

76.0 ± 1.4

75°C: 6.9c

n.o.

slGGGPS

Bacteria Spirosoma linguale

Hexamer

20

n.o.

83.1 ± 0.9

75°C: 16.4c

n.o.

fjGGGPS

Bacteria Flavobacterium johnsoniae

Dimer

20

54.1 ± 0.1

110.0 ± 0.1

50°C: 5.3c

242

a

mtGGGPS serves as reference, DSC and tapp1/2 data are from Table 4. For clarity, T1* of mtGGGPS is not shown. b T1/T2: apparent midpoint temperature, thermal transitions. T1 refers to the first observable transition (partial denaturation), T2 refers to the last observable transition (complete denaturation of the protein), average values of duplicates were calculated (given with standard deviations). n.o.; no transition corresponding to a partial denaturation observed c tapp1/2 was obtained by fitting a single exponential decay equation to the data from kinetic thermal inactivation studies (Figures S10, S14). d Due to high stability, a tapp1/2 value could not be determined even at 100 °C; after 120 min of incubation, 86.5 % (+/-5.1 %) of initial activity were remaining. e The enthalpy change of the first transition of unfolding was calculated as the area under the first transition peak from the baseline-corrected DSC data shown in Figure S13A and is given per mol of subunit. n.o.; no transition corresponding to a partial denaturation observed. * 0.1 mM GdmCl was added before heat incubation

Three central secondary structure elements of mtGGGPS possess high structural flexibility at elevated temperatures The findings presented above suggest that the first transition during thermal denaturation, which leads to catalytic inactivation, is due to the partial unfolding of mtGGGPS. We used Molecular Dynamics (MD) simulations to identify flexible structural elements that might tend to disintegrate at elevated temperatures. However, studying the dynamics of the full hexamer via

31 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 41

MD simulations within an acceptable time interval was not feasible due to the size of the complex. Thus, we focused on the analysis of one subunit. In order to estimate local flexibility, we determined the residue-specific root mean square fluctuation (RMSF), which is a measure of movements observed in a series of MD snapshots. The monomeric mtGGGPS variants have their first transition near 50 °C. We reasoned that MD simulations performed at 25 °C and 45 °C can serve as reference for MD simulations at 80 °C, which significantly exceeds the first transition temperature. Indeed, the resulting RMSF values are in agreement with these expectations: The amplitudes in the simulation at 80 °C are noticeable larger than in the reference simulations at 25 °C and 45 °C, most of the reference values are below 3Å (Figure S15). To indicate the flexibility of residues in detail, RMSF values were color coded and mapped onto the mtGGGPS_wt structure (Figure 12A). Only at 80 °C, strikingly high RMSF values beyond 3Å are accumulated in four 2D elements, namely helix α4, helix α5´, loop α6, and helix α8. For a statistical assessment, we deduced dev(C ,i ) values indicating for the C atom of each residue i a z-score corresponding to the deviation of the RMSFTC80 ,i value from the reference values C C RMSFT25 ,i and RMSFT45 ,i (Figure 12B). The mean dev(C ,i ) values of the mentioned 2D

elements are significantly higher than in any other structural motif within the protein. Three of the four identified structural elements contribute to the substrate binding sites of mtGGGPS. Helix α4 covers the long hydrophobic binding pocket for the GGPP9,

24, 25

. The loop α6

participates in G1P binding9. The dimeric variant mtGGGPS_W141A shows a drastically increased KM value for G1P (Table 2). This indicates that this loop, which is located close to the ring interface might be stabilized at increased temperatures by hexamerization. Most notably is helix α5´, which covers the GGPP binding site like helix α4. It harbors the “aromatic anchor“ W141 and is part of the ring interface, and therefore certainly receives extra stabilization by hexamerization. In contrast, the flexibility of the C-terminal residues constituting α-helix 8 might be an artifact caused by the analysis of an isolated monomer.

32 ACS Paragon Plus Environment

Page 33 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 12. MD simulation analysis of one subunit of the hexameric mtGGGPS_wt. (A) MD simulations were performed for three different temperatures, namely T = 25 °C, 45 °C, and 80 °C. Residue-specific RMSF values were calculated and mapped onto the initial structure. The color code is given by the bar on the right, all values above 3 Å are shown in red. (B) Plot of zscores indicating the deviation of residue-specific RMSF values (T = 80 °C) from reference values (T = 25 °C, 45 °C). The locations of helix α4, helix α5´, loop α6 and of helix α8 are indicated in both panels.

33 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 41

DISCUSSION We have analyzed the hexameric structure of mtGGGPS in detail and generated a number of variants with different oligomerization state. As described previously9, the hexamer can be regarded as a trimer of dimers, with dimer modules of the same topology as seen in native dimeric GGGPS. Besides the “dimer module interface”, two additional interfaces exist in the hexamer, which we call “interconnecting interface” and “ring interface”. The ring interface shows the highest level of conservation, but the interconnecting interface is equally important for the stabilization of the hexamer. For both interfaces, introduced mutations lead to a disruption of the hexamer. In contrast, the dimer module interface is obviously less important for hexamer stabilization, because mutations introduced here do not destabilize the structure in the hexameric context. Since many years, oligomerization is discussed as a key factor to increase the thermostability of proteins13, 14. A number of studies demonstrate that variants with an artificially reduced degree of oligomerization lose their activity at significantly lower temperatures than wild type proteins.27-29 Only a few publications go beyond and elucidate in more detail, how oligomerization enhances thermostability on the molecular level. For example, intersubunit ion pairs that form upon oligomerization have been shown to mediate higher thermostability in some proteins.30, 31 The thermal stability and unfolding characteristics of GGGPS somewhat resemble those of Lisoaspartyl-O-methyltransferase from Sulfolobus tokodaii.32 This protein forms a hexamer as well, and mutations have been introduced to disrupt the complex to dimers and monomers. Like for mtGGGPS, these variants are less thermostable, as shown by DSC and CD, but all variants show only single transitions that lead to denaturation. In contrast, all variants of mtGGGPS and even dimeric fjGGGPS revealed a biphasic thermal denaturation, with a first transition that is almost silent in CD spectroscopy, but leads to a loss of activity. While many (βα)8-barrel proteins denature cooperatively and reversibly in a single step, non-two-state denaturation profiles have occasionally been observed.33,

34

The GdmCl-induced denaturation of glycosomal TIM from

Trypanosoma brucei even involves two intermediates, with a first transition that is only slightly visible in CD spectroscopy. This denaturation step putatively results in a residual (βα)4β5structure, but leads to a loss of substrate binding capabilities.35 Other well examined examples of proteins with a non-two-state nature of thermal denaturation are the harpin protein HrpZ36 and

34 ACS Paragon Plus Environment

Page 35 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

GrpE from Thermus thermophilus37. In both cases, the first transitions (according to DSC experiments) are invisible in CD spectroscopy. In HrpZ, the first transition only reflects the dissociation of the native oligomeric protein complex to dimers. In GrpE however, the first transition represents the unfolding of the C-terminal β-sheet domain, while the second transition is correlated with the complete unfolding of the protein. The extremely high second transition temperature observed in our study suggest that the core of the GGGPS (βα)8-barrel is highly thermostable. CD spectroscopy and MD simulations indicate that only small sections of the protein fold, putatively close to the active site, collapse in the course of the first transition that is observable in DSC experiments. Interestingly, the temperature of the first transition is dependent on the quaternary structure, and rises with the complexity of the oligomerization state. This indicates that small parts of the protein benefit from extra stabilization by oligomerization to sustain activity at high temperatures. Hyperthermostable proteins have to deal with a tradeoff between thermostability and flexibility.13, 38 To be stable, proteins need to be tightly packed and to maximize internal interactions. To be active, enzymes need to be flexible to allow fast substrate binding, catalysis and product release. The combination of an intrinsically stable fold with an extra stabilization of critical protein sections by oligomerization might allow proteins to retain more flexibility on the protomer level, which is a prerequisite for high catalytic efficiency. The example of GGGPS thereby adds a novel aspect to the benefits of oligomerization for protein thermostability.

35 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 41

ASSOCIATED CONTENT Supporting Information. A table of all used protein variants with expression yields (Table S1), a figure showing the topology of the GGGPS fold (Figure S1), SDS PAGE demonstrating the purity of the expressed proteins (Figure S2), CD spectra of all mtGGGPS variants (Figure S3), supporting SEC data for determining oligomerization states (Figures S4, S5, S8), supporting DSC and nanoDSF data on thermostability of mtGGGPS variants and fjGGGPS (Figures S6, S7, S13), activity assays of all mtGGGPS variants (Figures S9, S12), thermal inactivation kinetics of mtGGGPS variants and fjGGGPS (Figures S10, S11, S14), and supporting MD data (Figure S15).

AUTHOR INFORMATION Corresponding Author * Institute of Biophysics and Physical Biochemistry, University of Regensburg, 93040 Regensburg, Germany, Tel.: +49 941 943 1634; E-mail: [email protected] Author Contributions ML did all the experimental work and drafted the manuscript. KH performed MD simulations, RM and RS assisted with advice and revised the manuscript. PB conceived of the study, coordinated experiments, and wrote the manuscript. All authors read and approved the final manuscript. Funding Sources This work was supported by a research grant from the Deutsche Forschungsgemeinschaft (BA 3943/2-1). ACKNOWLEDGMENT We thank 2bind GmbH for access to the Prometheus NT.48 instrument (NanoTemper Technologies) and support in nanoDSF experiments. We are grateful to Whitney Kilu, Cosimo Kropp and Sabine Laberer for technical assistance.

36 ACS Paragon Plus Environment

Page 37 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

ABBREVIATIONS CD, Circular dichroism; DSC, Differential scanning calorimetry; fjGGGPS, Flavobacterium johnsoniae GGGPS; GGPP, geranylgeranyl pyrophosphate; GGGP, geranylgeranylglyceryl phosphate; GGGPS, geranylgeranylglyceryl phosphate synthase; G1P, glycerol 1-phosphate; G3P, glycerol 3-phosphate; IPTG, isopropyl-β-D-1-thiogalactopyranoside; MD simulations, Molecular dynamics simulations; mtGGGPS, Methanothermobacter thermautotrophicus GGGPS; nanoDSF, Differential scanning fluorimetry; PNPase, bacterial purine nucleoside phosphorylase; PPase, Escherichia coli pyrophosphatase; RMSF, root mean square fluctuation; SEC, size exclusion chromatography; XOD, microbial xanthine oxidase

REFERENCES

[1] Kates, M. (1993) Membrane lipids in Archaea, In The biochemistry of archaea (Kates, M., Kushner, D. J., and Matheson, A. T., Eds.), pp 261-295, Elsevier, Amsterdam. [2] Koga, Y., Nishihara, M., Morii, H., and Akagawa-Matsushita, M. (1993) Ether polar lipids of methanogenic bacteria: structures, comparative aspects, and biosyntheses, Microbiol. Rev. 57, 164-182. [3] Wächtershäuser, G. (2003) From pre-cells to Eukarya--a tale of two lipids, Mol. Microbiol. 47, 13-22. [4] Payandeh, J., and Pai, E. F. (2007) Enzyme-driven speciation: crystallizing Archaea via lipid capture, J Mol Evol 64, 364-374. [5] Glansdorff, N., Xu, Y., and Labedan, B. (2008) The Last Universal Common Ancestor: emergence, constitution and genetic legacy of an elusive forerunner, Biol. Direct 3, 29. [6] Lombard, J., Lopez-Garcia, P., and Moreira, D. (2012) Phylogenomic investigation of phospholipid synthesis in archaea, Archaea 2012, 630910. [7] Lombard, J., Lopez-Garcia, P., and Moreira, D. (2012) The early evolution of lipid membranes and the three domains of life, Nat. Rev. Microbiol. 10, 507-515. [8] Jain, S., Caforio, A., and Driessen, A. J. (2014) Biosynthesis of archaeal membrane ether lipids, Front. Microbiol. 5, 641.

37 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 41

[9] Peterhoff, D., Beer, B., Rajendran, C., Kumpula, E. P., Kapetaniou, E., Guldan, H., Wierenga, R. K., Sterner, R., and Babinger, P. (2014) A comprehensive analysis of the geranylgeranylglyceryl phosphate synthase enzyme family identifies novel members and reveals mechanisms of substrate specificity and quaternary structure organization, Mol. Microbiol. 92, 885-899. [10] Guldan, H., Matysik, F. M., Bocola, M., Sterner, R., and Babinger, P. (2011) Functional assignment of an enzyme that catalyzes the synthesis of an Archaea-type ether lipid in Bacteria, Angew. Chem. In.t Ed. Engl. 50, 8188-8191. [11] Guldan, H., Sterner, R., and Babinger, P. (2008) Identification and characterization of a bacterial glycerol-1-phosphate dehydrogenase: Ni2+-dependent AraM from Bacillus subtilis, Biochemistry 47, 7376-7384. [12] Linde, M., Peterhoff, D., Sterner, R., and Babinger, P. (2016) Identification and characterization of heptaprenylglyceryl phosphate processing enzymes in Bacillus subtilis, J. Biol. Chem. 291, 14861-14870. [13] Sterner, R., and Liebl, W. (2001) Thermophilic adaptation of proteins, Crit. Rev. Biochem. Mol. Biol. 36, 39-106. [14] Vieille, C., and Zeikus, G. J. (2001) Hyperthermophilic enzymes: sources, uses, and molecular mechanisms for thermostability, Microbiol. Mol. Biol. Rev. 65, 1-43. [15] Zheng, L., Baumann, U., and Reymond, J. L. (2004) An efficient one-step site-directed and site-saturation mutagenesis protocol, Nucleic Acids Res. 32, e115. [16] Gasteiger, E., Hoogland, C., Gattiker, A., Duvaud, S., Wilkins, M. R., Appel, R. D., and Bairoch, A. (2005) Protein identification and analysis tools on the ExPASy server, In The proteomics protocols handbook (Walker, J. M., Ed.), pp xviii, 988 p., Humana Press, Totowa, N.J. [17] Bligh, E. G., and Dyer, W. J. (1959) A rapid method of total lipid extraction and purification, Can. J. Biochem. Physiol. 37, 911-917. [18] Kates, M. (1986) Techniques of lipidology: isolation, analysis, and identification of lipids, In Laboratory techniques in biochemistry and molecular biology vol. 3, part 2 (Burdon, R. H., and van Knippenberg, P. H., Eds.) 2nd rev. ed., pp 1-464, Elsevier, Amsterdam. [19] Suarez, A. S. G., Stefan, A., Lemma, S., Conte, E., and Hochkoeppler, A. (2012) Continuous enzyme-coupled assay of phosphate- or pyrophosphate-releasing enzymes, Biotechniques 53, 99-103. [20] Good, N. E., Winget, G. D., Winter, W., Connolly, T. N., Izawa, S., and Singh, R. M. (1966) Hydrogen ion buffers for biological research, Biochemistry 5, 467-477.

38 ACS Paragon Plus Environment

Page 39 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

[21] Krieger, E., Darden, T., Nabuurs, S. B., Finkelstein, A., and Vriend, G. (2004) Making optimal use of empirical energy functions: force-field parameterization in crystal space, Proteins 57, 678-683. [22] Schrodinger, LLC. (2015) The PyMOL Molecular Graphics System, Version 1.8. [23] Henn-Sax, M., Hocker, B., Wilmanns, M., and Sterner, R. (2001) Divergent evolution of (βα)8-barrel enzymes, Biol. Chem. 382, 1315-1320. [24] Payandeh, J., Fujihashi, M., Gillon, W., and Pai, E. F. (2006) The crystal structure of (S)-3O-geranylgeranylglyceryl phosphate synthase reveals an ancient fold for an ancient enzyme, J. Biol. Chem. 281, 6070-6078. [25] Peterhoff, D., Zellner, H., Guldan, H., Merkl, R., Sterner, R., and Babinger, P. (2012) Dimerization determines substrate specificity of a bacterial prenyltransferase, Chembiochem 13, 1297-1303. [26] Soderberg, T., Chen, A., and Poulter, C. D. (2001) Geranylgeranylglyceryl phosphate synthase. Characterization of the recombinant enzyme from Methanobacterium thermoautotrophicum, Biochemistry 40, 14847-14854. [27] Thoma, R., Hennig, M., Sterner, R., and Kirschner, K. (2000) Structure and function of mutationally generated monomers of dimeric phosphoribosylanthranilate isomerase from Thermotoga maritima, Structure 8, 265-276. [28] Schwab, T., Skegro, D., Mayans, O., and Sterner, R. (2008) A rationally designed monomeric variant of anthranilate phosphoribosyltransferase from Sulfolobus solfataricus is as active as the dimeric wild-type enzyme but less thermostable, J. Mol. Biol. 376, 506-516. [29] Clantin, B., Tricot, C., Lonhienne, T., Stalon, V., and Villeret, V. (2001) Probing the role of oligomerization in the high thermal stability of Pyrococcus furiosus ornithine carbamoyltransferase by site-specific mutants, Eur. J. Biochem. 268, 3937-3942. [30] Rahman, R. N., Fujiwara, S., Nakamura, H., Takagi, M., and Imanaka, T. (1998) Ion pairs involved in maintaining a thermostable structure of glutamate dehydrogenase from a hyperthermophilic archaeon, Biochem. Biophys. Res. Commun. 248, 920-926. [31] Vetriani, C., Maeder, D. L., Tolliday, N., Yip, K. S., Stillman, T. J., Britton, K. L., Rice, D. W., Klump, H. H., and Robb, F. T. (1998) Protein thermostability above 100°C: a key role for ionic interactions, Proc. Natl. Acad. Sci. U.S.A. 95, 12300-12305. [32] Tanaka, Y., Tsumoto, K., Yasutake, Y., Umetsu, M., Yao, M., Fukada, H., Tanaka, I., and Kumagai, I. (2004) How oligomerization contributes to the thermostability of an archaeon protein. Protein L-isoaspartyl-O-methyltransferase from Sulfolobus tokodaii, J. Biol. Chem. 279, 32957-32967.

39 ACS Paragon Plus Environment

Biochemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 41

[33] Labastida-Polito, A., Garza-Ramos, G., Camarillo-Cadena, M., Zubillaga, R. A., and Hernandez-Arana, A. (2015) Complex kinetics and residual structure in the thermal unfolding of yeast triosephosphate isomerase, BMC Biochem. 16, 20. [34] Carstensen, L., Zoldak, G., Schmid, F. X., and Sterner, R. (2012) Folding mechanism of an extremely thermostable (βα)8-barrel enzyme: a high kinetic barrier protects the protein from denaturation, Biochemistry 51, 3420-3432. [35] Guzman-Luna, V., and Garza-Ramos, G. (2012) The folding pathway of glycosomal triosephosphate isomerase: structural insights into equilibrium intermediates, Proteins 80, 1669-1682. [36] Tarafdar, P. K., Vedantam, L. V., Sankhala, R. S., Purushotham, P., Podile, A. R., and Swamy, M. J. (2014) Oligomerization, conformational stability and thermal unfolding of Harpin, HrpZPss and its hypersensitive response-inducing c-terminal fragment, C-214HrpZPss, PLoS One 9, e109871. [37] Groemping, Y., and Reinstein, J. (2001) Folding properties of the nucleotide exchange factor GrpE from Thermus thermophilus: GrpE is a thermosensor that mediates heat shock response, J. Mol. Biol. 314, 167-178. [38] Karshikoff, A., Nilsson, L., and Ladenstein, R. (2015) Rigidity versus flexibility: the dilemma of understanding protein thermal stability, FEBS J. 282, 3899-3917.

40 ACS Paragon Plus Environment

Page 41 of 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

TOC GRAPHIC For Table of Contents Use Only

41 ACS Paragon Plus Environment