The high‐resolution structure of a UDP‐L‐rhamnose synthase from Acanthamoeba polyphaga Mimivirus

Abstract

For the field of virology, perhaps one of the most paradigm‐shifting events so far in the 21st century was the identification of the giant double‐stranded DNA virus that infects amoebae. Remarkably, this virus, known as Mimivirus, has a genome that encodes for nearly 1,000 proteins, some of which are involved in the biosynthesis of unusual sugars. Indeed, the virus is coated by a layer of glycosylated fibers that contain d‐glucose, N‐acetyl‐d‐glucosamine, l‐rhamnose, and 4‐amino‐4,6‐dideoxy‐d‐glucose. Here we describe a combined structural and enzymological investigation of the protein encoded by the open‐reading frame L780, which corresponds to an l‐rhamnose synthase. The structure of the L780/NADP⁺/UDP‐l‐rhamnose ternary complex was determined to 1.45 Å resolution and refined to an overall R‐factor of 19.9%. Each subunit of the dimeric protein adopts a bilobal‐shaped appearance with the N‐terminal domain harboring the dinucleotide‐binding site and the C‐terminal domain positioning the UDP‐sugar into the active site. The overall molecular architecture of L780 places it into the short‐chain dehydrogenase/reductase superfamily. Kinetic analyses indicate that the enzyme can function on either UDP‐ and dTDP‐sugars but displays a higher catalytic efficiency with the UDP‐linked substrate. Site‐directed mutagenesis experiments suggest that both Cys 108 and Lys 175 play key roles in catalysis. This structure represents the first model of a viral UDP‐l‐rhamnose synthase and provides new details into these fascinating enzymes.

Short abstract

PDB Code(s): 7JID;

Abbreviations

CHES

2‐(cyclohexylamino)ethanesulfonic acid

DMSO

dimethyl sulfoxide

dTDP

thymidine diphosphate

HEPES

N‐2‐hydroxyethylpiperazie‐N′‐2‐ethanesulfonic acid

HEPPS

N‐2‐hydroxyethylpiperazine‐N′‐3‐propanesulfonic acid

HPLC

high‐performance liquid chromatography

NADP⁺

nicotinamide adenine dinucleotide phosphate (oxidized)

NADPH

nicotinamide adenine dinucleotide phosphate (reduced)

Ni‐NTA

nickel‐nitrilotriacetic acid

TEV

Tobacco Etch Virus

Tris

tris‐(hydroxymethyl)aminomethane

UDP

uridine diphosphate

1. INTRODUCTION

Within the last few decades there has been a decided emphasis placed on “translational research” or “from bench to bedside.” Along with this push, the term “result” has often been replaced with “deliverable.” Yet it cannot be denied that the reason why the HIV protease inhibitors were developed so rapidly and approved by the Federal Drug Administration beginning in 1995 was because of the previous decades of basic research on retroviruses and the aspartyl proteases. Indeed, the field of virology has been a major driving force toward understanding basic biochemical and cellular functions. A simple perusal of the literature shows the importance of viral research in elucidating such fundamental processes as DNA replication and transcription and the regulation of gene expression. ¹ And given the current deadly pandemic resulting from the coronavirus, SARS‐CoV‐2, it is clear that the field of virology is, by no means, “retro,” and that curiosity‐based research is still not outdated.

As recently as 2003, the virology community was rocked by the paradigm‐shifting discovery of a giant double‐stranded DNA virus that infects amoebae. ² This virus, known as Acanthamoeba polyphaga Mimivirus or simply Mimivirus, is larger than some bacteria where it is visible under a light microscope. Indeed, it was first misidentified as a Gram‐positive organism in 1992. ³ Over the last 16 years, nine additional giant viruses have been discovered. ⁴

The Mimivirus, with an overall diameter of ~750 μm, is coated by a layer of glycosylated fibers that contain d‐glucose, N‐acetyl‐d‐glucosamine, l‐rhamnose, and 4‐amino‐4,6‐dideoxy‐d‐glucose. ⁵ , ⁶ , ⁷ This heavily glycosylated layer is required for adhesion to the host organism. ⁸ Although it is not known that giant viruses are capable of widespread human infections, there is recent evidence to suggest that they may be responsible for some forms of pneumonia. ⁹ , ¹⁰

Strikingly, the genome of the Mimivirus contains 1,262 putative open reading frames. ¹¹ Two of these open reading frames encode for a sugar 4,6‐dehydratase (R141) and an l‐rhamnose synthase (L780), also known as 3,5‐epimerase/4‐reductase. ¹² The reactions catalyzed by these two enzymes, as highlighted in Scheme 1, are required for the biosynthesis of l‐rhamnose. Note that the production of this 6‐deoxy sugar proceeds via a UDP‐ or dTDP‐sugar intermediate depending upon the organism under investigation.

SCHEME 1.

Pathway for the biosynthesis of UDP‐l‐rhamnose

l‐rhamnose, observed in plants, bacteria, fungi, and viruses, has been isolated from structural polysaccharides as well as found attached to glycoproteins and natural products. In some cases, l‐rhamnose has been implicated in pathogenesis. ¹³ Like that observed for the production of l‐rhamnose in Mimivirus, plants and fungi also contain a bifunctional L780 equivalent with both 3,5‐epimerization and 4‐reduction activities. ¹⁴ , ¹⁵ In bacteria, the 3,5‐epimerase and 4‐reductase activities are found on separate polypeptide chains. ¹⁶

We recently reported the X‐ray structure of R141 from Mimivirus to 2.05 Å resolution. ¹⁷ Our results demonstrated several significant architectural differences between the Mimivirus 4,6‐dehydratase and those previously characterized bacterial enzymes. As a continuation of our studies on sugar‐modifying enzymes from Mimivirus, herein we report a structural and enzymological characterization of L780. The structure, determined to a nominal resolution of 1.45 Å, places L780 into the bifunctional family of short‐chain dehydrogenase/reductases, which also includes the GDP‐fucose synthases. Importantly, the model presented is that of a ternary complex with NADP⁺ and UDP‐l‐rhamnose (the product), and thus it provides new insight into the bifunctional UDP‐l‐rhamnose synthases.

2. RESULTS

2.1. The structure of L780

The crystals utilized in this investigation were grown in the presence of NADP⁺ and UDP‐l‐rhamnose (the product of L780 catalysis). They belonged to the space group P1 with two subunits in the asymmetric unit. Gel filtration experiments demonstrated that the protein migrates as a dimer. The model was refined to a nominal resolution of 1.45 Å with an overall R‐factor of 19.9%.

Shown in Figure 1a is a ribbon representation of the dimer. The overall fold of L780 places it into the well‐characterized short‐chain dehydrogenase/reductase superfamily. ¹⁹ The α‐carbons for the two subunits superimpose with a root‐mean‐square deviation of 0.32 Å. Subunits 1 and 2 extend from Met 1 to Gln 288 and Met 1 to Met 285, respectively, without breaks in the polypeptide chains. Each subunit contains a six‐stranded parallel β‐sheet and 11 α‐helices. The fourth and fifth α‐helices are primarily responsible for the formation of the subunit:subunit interface, which is distinctly hydrophobic in nature. Shown in Figure 1b is a close‐up stereo view of the striking interactions observed at the subunit:subunit interface. These include a parallel stacking interaction between Phe 92 (subunit 1) and Phe 149 (subunit 2), and a classical π‐cation interaction between the side chains of Tyr 85 (subunit 1) and Arg 145 (subunit 2) which is further augmented with a T‐shape stacking interaction between Tyr 85 (subunit 1) and Phe 142 (subunit 2). Located at approximately midway along the subunit:subunit interface is a leucine‐rich patch formed by Leu 84, Leu 88, and Leu 146 in subunit 1 and the twofold related residues in subunit 2. Leu 76, Val 77, Val 80, Val 138, and Val 139 in subunit 1 form an additional hydrophobic patch that lies against the symmetry‐related side chains in subunit 2. The total buried surface area of the dimer is ~2,100 Ų.

FIGURE 1.

Overall molecular architecture of L780. A ribbon representation is shown in (a) with the position of the local twofold rotational axis indicated by the arrow. Subunits 1 and 2 in the dimer are displayed in light blue and pink, respectively. The subunit:subunit interface, presented as a stereo view in (b), is dominated by hydrophobic interactions. This figure and Figures 2, 3, and 5 were prepared with PyMOL ¹⁸

The observed electron densities corresponding to the bound ligands, NADP⁺ and UDP‐l‐rhamnose in subunit 1, are displayed in Figure 2. The electron density for the dinucleotide shows that the NADP⁺ had been hydrolyzed to ADP‐ribose 2′‐phosphate. This phenomenon has been previously observed in several of our structural investigations on dinucleotide‐dependent enzymes, and it is most likely an artifact of the experiment. ¹⁷ , ²⁰ The electron density for the dinucleotide in subunit 2 also indicated that the nicotinamide ring had been hydrolyzed from the cofactor. The electron density corresponding to UDP‐l‐rhamnose was unambiguous in both subunits, however (Figure 2).

FIGURE 2.

Observed electron density for the bound ligands in subunit 1. The electron density map, shown in stereo, was calculated with (F _o − F _c) coefficients and contoured at 3σ. The ligands were not included in the X‐ray coordinate file used to calculate the omit map, and thus there is no model bias

Close‐up views of the regions surrounding the NADP⁺ and UDP‐sugar ligands are displayed in Figure 3a,b, respectively. As can be seen in Figure 3a, the adenine ring of NADP⁺ is anchored into the active site by the side chains of Asp 35 and Asp 82, the backbone amide of Ala 34, and a water molecule. Additionally, the guanidinium moiety of Arg 33 participates in a π‐cation interaction with the adenine ring. Both Arg 33 and Arg 58 participate in electrostatic interactions with the phosphoryl group attached to the ribose 2‐hydroxyl. Five water molecules surround the phosphoryl and ribosyl groups. There are eight additional water molecules lying within 3.2 Å of the dinucleotide. The only side chain that directly interacts with the pyrophosphoryl group of the cofactor is contributed by His 60. The backbone amide groups of Trp 11 and Ile 12 also form hydrogen‐bonding interactions with the pyrophosphoryl oxygens. Members of the short‐chain dehydrogenase/reductase superfamily typically contain the characteristic signature sequence Tyr‐X‐X‐X‐Lys. In L780, these residues correspond to Tyr 136 and Lys 140. The position of these residues can be seen in both Figure 3a,b.

FIGURE 3.

Close‐up views of the active site in subunit 1. Presented in stereo in (a) and (b) are the active site regions surrounding the NADP⁺ and UDP‐l‐rhamnose ligands, respectively. Ordered water molecules are displayed in red spheres. The dashed lines indicate possible hydrogen‐bonding interactions within 3.2 Å

With respect to the UDP‐l‐rhamnose ligand, the uracil ring forms a stacking interaction with the side chain of Tyr 179, which additionally hydrogen bonds to an α‐phosphoryl oxygen. Both the carbonyl group of Arg 181 and the backbone amide of Cys 183 lie within 3.2 Å of the uracil base. It is further surrounded by two water molecules. The ribosyl group is enveloped by the side chains of His 221 and Gln 246, and two water molecules. In addition to the interaction with Tyr 179, the α‐phosphoryl oxygens also form electrostatic interactions with the side chains of Lys 175 and Arg 254. Four water molecules lie within 3.2 Å of the pyrophosphoryl oxygens. Two water molecules lie within hydrogen‐bonding distance to the C‐2′ hydroxyl group. The C‐3′ hydroxyl of the l‐rhamnose moiety is situated within 2.9 Å of the backbone amide of Met 161. The side chains of Thr 106 and Tyr 136, along with a water molecule, anchor the C‐4′ hydroxyl group into the active site cleft.

Similar to what has been shown in the studies on GDP‐fucose synthase, Cys 108 in L780 can be predicted to serve as the enzymatic base required for the deprotonation of the substrate at the C‐3′ carbon subsequently leading to an enolate intermediate. ²¹ In the L780/NADP⁺/product complex described here, S^γ of Cys 108 lies at 3.5 Å from this carbon (Figure 3b). Likewise, again based on that observed for GDP‐fucose synthase, it is possible that Lys 175 in L780 functions as an enzymatic acid by protonating the opposite face. In the L780/NADP⁺/product complex, N^ζ of the side chain lies at 2.8 and 3.1 Å, respectively, from the bridging oxygen of the pyrophosphoryl group and an α‐phosphoryl oxygen, respectively. It is, however, 6.9 Å from the C‐3′ carbon. This will be discussed further below.

2.2. Kinetic properties of L780

Steady‐state kinetic parameters for L780 were determined via a spectrophotometric assay. Plots of reaction rates versus substrate concentrations are displayed in Figure 4. The k _cat and K _M values, determined for UDP‐4‐keto‐6‐deoxy‐d‐glucose, dTDP‐4‐keto‐6‐deoxy‐d‐glucose, and NADPH are listed in Table 2. The measured value for K _M with UDP‐4‐keto‐6‐deoxy‐d‐glucose was similar to that previously determined for L780. ¹² The catalytic efficiencies of L780 using UDP‐4‐keto‐6‐deoxy‐d‐glucose and dTDP‐4‐keto‐6‐deoxy‐d‐glucose as substrates were 3.3 × 10³ (M⁻¹ s⁻¹) and 7.6 × 10² (M⁻¹ s⁻¹), respectively. Although L780 prefers the UDP‐sugar substrate, it can still function on the dTDP‐linked version.

FIGURE 4.

Plots of initial velocities versus substrate concentrations. In presenting the data as we do, we are adhering to standard conventions in enzymology. Measuring velocities over a wide range of substrate concentrations allows us to obtain data that define both k _cat and k _cat/K _M well, which is not accomplished by measuring replicates at fewer different concentrations. The graphs shown allow for a qualitative appreciation of the quality of the data; the quantitative goodness‐of‐fit to the Michaelis–Menten equation is given by the standard errors as described in Section 4. Those plots displayed in (a), (b), and (c) are for the wild‐type enzyme. That presented in (d) is for the K165A variant

TABLE 2.

Kinetic parameters for L780

Protein	Substrate	K M (mM)	k cat (s−1)	k cat/K M (M−1 s−1)
L780	UDP‐4‐keto‐6‐deoxy‐d‐glucose	0.26 ± 0.02	0.85 ± 0.05	3.3 (±0.4) × 103
	dTDP‐4‐keto‐6‐deoxy‐d‐glucose	0.63 ± 0.07	0.48 ± 0.06	7.6 (±0.5) × 102
	NADPH	0.011 ± 0.001	1.03 ± 0.06	9.4 (±0.9) × 104
K175A variant	UDP‐4‐keto‐6‐deoxy‐d‐glucose	1.17 ± 0.19	0.021 ± 0.003	18 ± 3

In addition to measuring the kinetic parameters for the wild‐type enzyme, those for the following site‐directed mutant proteins were also evaluated: C108S, C108S/K175A, and K175A. The former two demonstrated no measurable activity. The K175A mutant protein was active but had a higher K _M and a significantly lower k _cat thereby leading to an overall catalytic activity of 18 ± 3 (M⁻¹ s⁻¹).

3. DISCUSSION

A search utilizing PDBeFOLD ²² revealed that L780 most closely matches the bifunctional epimerase/reductase from Arabidopsis thaliana. ²³ A sequence alignment for these two proteins is provided in Figure 5a. They demonstrate amino acid similarities and identities of 59 and 41%, respectively. An examination of the Protein Data Bank coordinates for the A. thaliana protein (4QQR) revealed some unusual structural features. The enzyme was crystallized from 1.8 M ammonium sulfate, and the X‐ray data were collected to 2.7 Å resolution. At such a salt concentration and at moderate resolution, one would expect few observed water molecules. In the deposited coordinates there were 68 water molecules listed. A chloride was shown to coordinate with a zinc ion, which seemed odd given that the coordination geometry of zinc ions in proteins is usually via oxygen, nitrogen, and sulfur atoms. ²⁴ Likewise, there were two highly unusual cis peptide bonds in subunit 1 between Gly 69 and Val 79 and between Pro 74 and Asn 75. To investigate these properties further, we utilized the deposited structure factors and re‐examined the model. After refinement of the model, it became clear that only nine waters were observable. The chloride is most likely a partially disordered water molecule. A comparison of the regions between Gly 69 and Val 76 in the two models is presented in Figure 5b. Admittedly, the electron density is weaker in this region but there is no chemical justification for building the protein backbone with cis conformations. In fact, the sulfate ion that was included in the PDB deposition is most likely the side chain of Thr 71. There are numerous incorrect rotamer conformations in the deposited model as well. Given that the high‐resolution limit of the X‐ray diffraction data was only 2.7 Å, there was no justification to build the side chains of Thr 182 and Thr 250 in eclipsed conformations. These residues can easily be modeled into the electron density in the staggered conformation that is more chemically probable. Removing the incorrectly placed water molecules and correcting for various modeling mistakes it was possible to lower the overall R‐factor from 20.5 to 19.5%. This is yet another cautionary tale concerning the quality of some X‐ray models deposited in the Protein Data Bank.

FIGURE 5.

Comparison of L780 to the UDP‐l‐rhamnose synthase from A. thaliana. An amino acid sequence alignment is presented in (a) with the positions of the β‐strands and the α‐helices indicated by the pink arrows and blue rectangles, respectively. The X‐ray coordinates deposited in the Protein Data Bank under the accession number 4QQR were re‐examined. Shown in stereo in (b) is the region where the structure deposited under 4QQR was not chemically probable. The original structure, which contained two cis peptide bonds, is displayed in purple, whereas our interpretation of the electron density is shown in the model highlighted in white bonds. A superposition of the ribbon drawings for L780 (purple) and the A. thaliana UDP‐l‐rhamnose synthase (light blue) is displayed in stereo in (c). The bound ligands, presented in stick representations, belong to the L780 model

Utilizing the re‐examined model, the α‐carbons for L780 and the A. thaliana protein superimpose with a root‐mean‐square deviation of 1.3 Å (subunit 1). A superposition of the two models is presented in Figure 5c. There is only one region, from Phe 55 to Lys 75 (L780 numbering), whereby the two proteins differ significantly. This difference is mostly likely due to the fact that the A. thaliana protein structure was determined in the absence of bound ligands. Indeed, this region in the A. thaliana apoprotein projects into the active site such that it would sterically clash if the dinucleotide and the nucleotide‐linked sugar were bound. Caution must be exercised when interpreting models devoid of relevant ligands. As an example, it was shown that the A. thaliana protein displays a comparable affinity for both NADH and NADPH, and it was reasoned that the additional phosphate group is solvent‐exposed and does not interact with the protein. ²³ In the L780 model, however, the phosphoryl group attached to the C‐2 hydroxyl ribose of NADP⁺ is anchored to the protein via the guanidinium groups of Arg 33 and Arg 58. The electron density in this region is unambiguous. Importantly, Arg 33 is conserved as Arg 45 in the A. thaliana protein (Figure 5a) but in the apoprotein it is projected toward the surface. Most likely, its position shifts upon dinucleotide binding.

The catalytic mechanism of L780, which requires two epimerizations at C‐3′ and C‐5′ and a reduction at C‐4′, is absolutely fascinating but not well understood. Previous studies on the GDP‐fucose synthases provide considerable insight, however. ²⁵ , ²⁶ , ²⁷ On the basis of the elegant study by Lau and Tanner, a mechanism for GDP‐fucose synthase has been put forth as shown in Scheme 2. ²¹ According to all the biochemical data, it is thought that Cys 109 functions as an active site base to abstract the C‐3′ proton leading to the formation of the first enolate intermediate. This intermediate is subsequently protonated by His 179, which serves as an active site acid. Following this first epimerization, there is a reset of the protonation states of Cys 109 and His 179 so that these residues can play analogous roles but this time functioning at the C‐5′ position. The manner in which this proton reset occurs is not known. In the final reduction step at C‐4′, the conserved Tyr 136 donates a proton to the C‐4′ keto oxygen and a hydride is transferred from NADPH to the C‐4′ carbon, a mechanism conserved among the short‐chain dehydrogenase/reductase superfamily.

SCHEME 2.

Proposed reaction mechanism for GDP‐l‐fucose synthase adapted from Lau and Tanner ²¹

In L780, the side chain of Cys 108 is positioned at ~3.5 Å from the C‐3′ carbon and thus could function as the active site base. In keeping with this structural observation, the C108S variant displayed no catalytic activity under the assay conditions utilized in this investigation. There is no corresponding histidine residue in L780, however. Rather there is a lysine, Lys 175, that could possibly play a similar role. In the current L780 model, Lys 175 interacts with the pyrophosphoryl group of the UDP‐sugar product, and its ε‐nitrogen is located at 5 and 6.4 Å from the pyranosyl C‐5′ carbon in subunits 1 and 2, respectively. There are two caveats with our model that must be recognized. The first is that the nicotinamide ring is missing from the dinucleotide. The second is that the model represents a ternary complex of the enzyme with ADP‐ribose 2′‐phosphate and UDP‐product, not NADPH and UDP‐substrate. Thus, it is reasonable to suggest that there is a movement of Lys 175 in the active site when the UDP‐sugar substrate binds to allow it to function as an active site acid. There are no other residues surrounding the pyranosyl moiety that could function in such capacity. Not surprisingly, the K175A variant displays a considerable reduction in its catalytic efficiency.

The model of L780 presented here represents the first structural analysis of a UDP‐l‐rhamnose synthase of viral origin. Given the current raging pandemic, it is clear that much remains to be learned about viruses in general. The fact that the Mimivirus was only discovered in 2003 suggests there is a wealth of new knowledge to be uncovered as long as curiosity‐based research is allowed to continue. These giant viruses may, indeed, be common in Earth's biosphere. Presently, it is unknown whether l‐rhamnose contributes to the pathogenicity of the Mimivirus with respect to its host, the amoeba. Importantly, giant viruses of amoebae have been detected in human samples. Although is not clear yet whether they may play a role in human disease, research directed toward their possible role is an emerging, exciting, and timely field. ²⁸

4. MATERIALS AND METHODS

4.1. Protein expression and purification of L780

The gene encoding L780 from A. polyphaga Mimivirus was synthesized by Integrated DNA Technologies and placed into pET28t3g, a modified pET28b(+) vector (Novagen), which leads to proteins with an N‐terminal polyhistidine tag as previously described. ²⁹ The pET28t3g‐L780 plasmid was utilized to transform Rosetta2(DE3) Escherichia coli cells (Novagen). Cultures were grown in lysogeny broth supplemented with kanamycin and chloramphenicol (both at 50 mg/L concentration) at 37°C with shaking until an optical density of 0.8 was reached at 600 nm. The flasks were cooled in an ice bath, and the cells were induced with 1.0 mM isopropyl β‐d‐1‐thiogalactopyranoside and allowed to express protein at 21°C for 24 h.

The cells were harvested by centrifugation and frozen as pellets in liquid nitrogen. These pellets were subsequently disrupted by sonication on ice in a lysis buffer composed of 50 mM sodium phosphate, 20 mM imidazole, 10% glycerol, and 300 mM NaCl (pH 8.0). The lysate was cleared by centrifugation, and the proteins were purified at 4°C utilizing Prometheus™ Ni‐NTA agarose (Prometheus Protein Biology Products) according to the manufacturer's instructions. All buffers were adjusted to pH 8.0 and contained 50 mM sodium phosphate, 300 mM NaCl, and imidazole concentrations of 20 mM for the wash buffer and 300 mM for the elution buffer. The polyhistidine tag was removed by digestion with TEV protease and the un‐cleaved L780 and the TEV protease were removed by passage over Ni‐NTA agarose. The enzyme was dialyzed into 10 mM Tris and 200 mM NaCl (pH 8.0) and concentrated to 20 mg/ml based on an extinction coefficient of 1.2 (mg/ml)⁻¹ cm⁻¹.

4.2. Site‐directed mutagenesis

The site‐directed variants, K185A, C108S, and C108S/K185A were prepared via the Stratagene QuikChange method. All variants were overexpressed, purified, and concentrated as described for the wild‐type enzyme.

4.3. Determination of kinetic constants

Steady‐state kinetic parameters for L780 were determined via a spectrophotometric assay. Reactions to obtain kinetic parameters for the NDP‐sugar substrates contained 50 mM HEPPS (pH 8.0), 0.4 mM NADPH, 0.027 mg/ml L780, and either UDP‐4‐keto‐6‐deoxy‐d‐glucose (0.01–4 mM) or dTDP‐4‐keto‐6‐deoxy‐d‐glucose (0.01–8 mM). Reactions were initiated by the addition of L780, and the decrease in absorbance was measured at 340 nm at 25°C on a Beckman Coulter DU‐640 spectrophotometer as the NADPH was converted to NADP⁺. Reactions to measure NADPH kinetic parameters contained 50 mM HEPPS (pH 8), 0.011 mg/ml L780, and 4 mM UDP‐4‐keto‐6‐deoxy‐d‐glucose. Reactions contained 0.001–0.2 mM NADPH and were again initiated by the addition of L780 and monitored at 340 nM.

Determination of the kinetic constants for the K175A enzyme was conducted in a similar manner to that described for the wild‐type enzyme using 0.88 mg/ml of the K175A variant and UDP‐4‐keto‐6‐deoxy‐d‐glucose concentrations ranging from 0.1 to 10 mM. No enzymatic activity was observed for the C108S and C108S/K175A variants.

The data were fitted to the equation: v ₀ = (V _max[S])/(K _M + [S]). The k _cat values were calculated according to the equation: k _cat = V _max/[E _T].

4.4. Synthesis and purification of UDP‐l‐rhamnose

A solution of 2.5 mM UDP‐d‐glucose, 3.0 mM NADPH, and 50 mM HEPPS was adjusted to pH 8.0. The R141 dehydratase, prepared as previously described, ¹⁷ was added to a final concentration of 0.5 mg/ml and L780 added to a final concentration of 1 mg/ml. The reaction was allowed to proceed overnight at room temperature. The enzymes were removed by filtration through a 30 kDa cutoff filter. The filtrate was then diluted 5× with water and loaded on a 50 ml HiLoad^tm 26/10 Q‐sepharose HP column (GE Healthcare). Reaction products were separated using a 13 column volume gradient from 0 to 1.2 M ammonium acetate at pH 4.0; the UDP‐rhamnose eluting at an ammonium acetate concentration of approximately 0.6 M.

4.5. Synthesis of either UDP‐4‐keto‐6‐deoxy‐d‐glucose or dTDP‐4‐keto‐6‐deoxy‐d‐glucose

50 mM solutions of either UDP‐d‐glucose or dTDP‐d‐glucose were prepared in 50 mM HEPPS (pH 8.0) containing 20 mg/ml of the R141 dehydratase. The reactions were allowed to proceed overnight at room temperature. After removal of the enzyme by filtration through a 30 kDa membrane, conversion to the 4‐keto sugar products was verified by HPLC.

4.6. Determination of the quaternary structure

The quaternary structure of L780 was evaluated using a Superdex 200 10/300 (GE Healthcare) gel filtration column with an ÄKTA purifier. The samples were loaded and run in 10 mM Tris (pH 8.0) and 200 mM NaCl at a speed of 0.5 ml/min. Standards run for comparison were bovine serum albumin (MW = 66,000) and carbonic anhydrase (MW = 29,000).

4.7. Crystallization

Crystals of L780 were grown via hanging drop vapor diffusion at room temperature from 20% poly(ethylene glycol) 8,000, 2% DMSO, and 100 mM CHES (pH 9.0) in the presence of 5 mM NADP⁺ and 5 mM UDP‐l‐rhamnose. They belonged to the triclinic space group P1 with unit cell dimensions of a = 49.8 Å, b = 53.0 Å, c = 70.1 Å, α = 81.9°, β = 88.7°, and γ = 66.7°. The asymmetric unit contained one dimer. For X‐ray data collection, the crystals were transferred to a cryo‐protectant solution composed of 28% poly(ethylene glycol) 8,000, 250 mM NaCl, 18% ethylene glycol, 5 mM NADP⁺, 5 mM UDP‐l‐rhamnose, and 100 mM HEPES (pH 7.5).

4.8. X‐ray data collection and processing

An X‐ray data set was collected using a BRUKER D8‐VENTURE sealed tube system equipped with HELIOS optics and a PHOTON II detector. The X‐ray data set was processed with SAINT and scaled with SADABS (Bruker AXS). Relevant X‐ray data collection statistics are listed in Table 2.

4.9. Structure solution and model refinement

L780 was solved via molecular replacement with Phaser ³⁰ using PDB entry 4QQR as a search probe. ³¹ Iterative cycles of model‐building with COOT ³² , ³³ and refinement with REFMAC ³⁴ led to a final X‐ray model with an overall R‐factor of 19.9% at 1.45 Å resolution. Relevant refinement statistics are provided in Table 1.

TABLE 1.

X‐ray data collection statistics and model refinement statistics

Resolution limits (Å)	50.0–1.45 (1.55–1.45) a
Number of independent reflections	110,458 (18,606)
Completeness (%)	99.5 (88.7)
Redundancy	4.1 (2.1)
Avg I/avg σ(I)	13.5 (2.7)
R sym (%) b	5.7 (38.4)
c R‐factor (overall)%/no. reflections	19.9/87647
R‐factor (working)%/no. reflections	19.8/104874
R‐factor (free)%/no. reflections	22.7/5584
Number of protein atoms	4,660
Number of heteroatoms	881
Average B values
Protein atoms (Å2)	18.0
Ligand (Å2)	19.7
Solvent (Å2)	29.5
Weighted RMS deviations from ideality
Bond lengths (Å)	0.011
Bond angles (°)	1.85
Planar groups (Å)	0.010
Ramachandran regions (%) d
Most favored	98.8
Additionally allowed	1.2
Generously allowed	0.0

^aStatistics for the highest resolution bin.

^b R _sym = (∑|I − $\overline{I}$|/∑I) × 100.

^c R‐factor = (∑|F _o − F _c|/∑|F _o|) × 100, where F _o is the observed structure‐factor amplitude and F _c ^. is the calculated structure‐factor amplitude.

^dDistribution of Ramachandran angles according to PROCHECK. ³⁵

AUTHOR CONTRIBUTIONS

Nicholas J. Bockhaus: Data curation. Justin D. Ferek: Investigation. James B. Thoden: Conceptualization; data curation; formal analysis; investigation; methodology. Hazel H. Holden: Conceptualization; funding acquisition; investigation; project administration; validation; visualization; writing‐review and editing.

CONFLICT OF INTEREST

The authors declare no potential conflict of interest.

Bockhaus NJ, Ferek JD, Thoden JB, Holden HM. The high‐resolution structure of a UDP‐L‐rhamnose synthase from Acanthamoeba polyphaga Mimivirus . Protein Science. 2020;29:2164–2174. 10.1002/pro.3928

DATA AVAILABILITY STATEMENT

X‐ray coordinates have been deposited in the Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, N. J. (accession no. 7JID).