Molecular architecture of KedS8, a sugar N-methyltransferase from Streptoalloteichus sp. ATCC 53650

Abstract

Kedarcidin, produced by Streptoalloteichus sp. ATCC 53650, is a fascinating chromoprotein of 114 amino acid residues that displays both antibiotic and anticancer activity. The chromophore responsible for its chemotherapeutic activity is an ansa-bridged enediyne with two attached sugars, l-mycarose, and l-kedarosamine. The biosynthesis of l-kedarosamine, a highly unusual trideoxysugar, is beginning to be revealed through bioinformatics approaches. One of the enzymes putatively involved in the production of this carbohydrate is referred to as KedS8. It has been proposed that KedS8 is an N-methyltransferase that utilizes S-adenosylmethionine as the methyl donor and a dTDP-linked C-4′ amino sugar as the substrate. Here we describe the three-dimensional architecture of KedS8 in complex with S-adenosylhomocysteine. The structure was solved to 2.0 Å resolution and refined to an overall R-factor of 17.1%. Unlike that observed for other sugar N-methyltransferases, KedS8 adopts a novel tetrameric quaternary structure due to the swapping of β-strands at the N-termini of its subunits. The structure presented here represents the first example of an N-methyltransferase that functions on C-4′ rather than C-3′ amino sugars.

Introduction

Kedarcidin, first isolated in 1991, is a chromoprotein antitumor antibiotic produced by Streptoalloteichus sp. ATCC 53650.^1,2 The apoprotein consists of 114 amino acid residues,² whereas the chromophore (Scheme 1), responsible for kedarcidin’s chemotherapeutic activity, belongs to the enediyne family of antitumor compounds. Attached to the chromophore is l-kedarosamine, a unique trideoxysugar whose biosynthesis has been largely uncharacterized. Indeed, only recently has a report appeared in the literature outlining a possible pathway for l-kedarosamine production starting from dTDP-glucose.³ As indicated in Scheme 1, the last step of l-kedarosamine biosynthesis is predicted to be the dimethylation of the C-4′ sugar amino group by the action of KedS8 and/or KedS9. Whereas KedS8 and KedS9 presumably employ S-adenosylmethionine (SAM) for activity, their enzymatic activities have not been established in vitro.

Figure 1. Scheme.

Reaction catalyzed by KedS8.

We became intrigued by the molecular architecture of KedS8 given our long-standing interest in SAM-dependent sugar methyltransferases. Indeed, the reports in the literature concerning the three-dimensional structures of the sugar N,N-dimethyltransferases, DesVI from Streptomyces venezuelae, and TylM1 from Streptomyces fradiae, have arisen from our research.^4–6 Both DesVI and TylM1 catalyze dimethylation reactions at the sugar C-3′ amino group. In l-kedarosamine, the amino group that is dimethylated is at the C-4′ position (Scheme 1). In the case of TylM1, it was possible trap its natural substrate, dTDP-3-amino-3,6-dideoxyglucose, into the active site.⁵ The model of TylM1 revealed that SAM and the dTDP-sugar were appropriately aligned for a direct in-line displacement reaction. In addition, site-directed mutagenesis experiments on TylM1 suggested that a catalytic base was not required to remove the proton from the amino group of the dTDP-sugar as the reaction proceeds. Most likely catalysis by TylM1 occurs via approximation with the proton from the sugar amino group presumably transferred to one of the water molecules lining the active site region.

Here we describe the three-dimensional architecture of KedS8 in complex with S-adenosylhomocysteine (SAH) determined to 2.0 Å resolution. Strikingly, due to an extended β-strand at the N-terminus, KedS8 adopts a unique quaternary structure.

Results and Discussion

The structure of the KedS8/SAH complex was solved to a nominal resolution of 2.0 Å and refined to an R_overall of 17.1%. KedS8 consists of 248 amino acid residues. Overall the electron density corresponding to the polypeptide chain backbone was well ordered and continuous from the N-terminus to Glu 245. It was somewhat weaker from Glu 15 to Gly 25.

The crystals used in the investigation belonged to the space group I222 with one monomer in the asymmetric unit. Typically the sugar N,N-methyltransferases function as dimers.^4–6 To explore the quaternary structure of KedS8 in solution, size exclusion chromatography experiments were conducted as described in Materials and Methods section. The data presented in Figure 2 are suggestive of KedS8 functioning as a tetramer.

Figure 2. Figure.

Analysis of the quaternary structure of KedS8 by gel filtration chromatography. Shown is retention time on the HPLC versus milliabsorbance units. KedS8 and TylM1 migrate as tetrameric and dimeric species, respectively (retention times of 14.5 and 16.8 min). Standards used for comparison: alcohol dehydrogenase (retention time 12.5 min, MW=150,000), albumin (retention time 16.3 min, MW=66,000), and carbonic anhydrase (retention time 17.9 min, MW=29,000).

Shown in Figure 3(a) is a stereo ribbon representation of the KedS8 tetramer as observed in the crystalline lattice. It has overall dimensions of ∼68 Å × 66 Å × 102 Å. The buried surface area for each subunit of the tetramer is ∼3,000 Ų. A close-up view of one subunit is displayed in Figure 3(b). The polypeptide chain of the subunit initiates with an extended β-strand followed by two α-helices connected by a Type I turn. Two major tertiary structural motifs dominate the fold of the subunit: a seven-stranded mixed β-sheet surrounded by four α-helices and a four-stranded antiparallel β-sheet. The active site is wedged between these two β-sheets.

Figure 3. Figure.

Structure of KedS8. Shown in (a) is a stereo ribbon representation of the KedS8 tetramer. A view of a single subunit is presented in (b) with the β-strands and α-helices highlighted in purple and green, respectively. The bound ligand, SAH, is shown in a stick representation. This figure and figures 3 and 4 were prepared using PyMOL.⁷

Electron density corresponding to SAH is presented in Figure 4(a). As can be seen, it is well ordered with the adenine ring in the anti conformation and the ribose adopting the C₂′-endo pucker. The adenine ring is held in place by hydrogen bonds provided by Asp 96 and a water molecule. In addition there are numerous hydrophobic interactions contributed by the side chains of Leu 75, Met 97, Phe 98, and Tyr 118. The carboxylate side chain of Glu 74 bridges the hydroxyl groups of the ribose. The hydroxyl of Thr 113 lies within hydrogen bonding distance of the carboxylate group of SAH. For the analysis reported here, the KedS8 crystals were soaked in a solution of 40 mM dTDP-benzene in an attempt to trap a substrate analogue into the active site. A similar ligand, UDP-benzene, was successfully utilized in our structural analysis of DesVI.⁴ Unfortunately, in the case of KedS8, there was no clear electron density for the dTDP-benzene ligand.

Figure 4. Figure.

Close-up view of the SAH binding pocket. The electron density corresponding to the bound SAH cofactor is shown. The map, contoured at 3σ, was calculated with coefficients of the form F_o–F_c, where F_o was the native structure factor amplitude and F_c was the calculated structure factor amplitude. The SAH ligand was not included in the coordinate file for the map calculation. Ordered water molecules are represented as red spheres. The dashed lines indicate possible hydrogen bonding interactions.

The fact that KedS8 behaves as a tetramer in solution was unexpected given our past experience with sugar N,N-dimethyltransferases. Shown in Figure 5(a) is one of the subunit:subunit interfaces of the tetramer. The four-stranded antiparallel β-sheet from one subunit abuts the other leading to a total buried surface area of 2300 Ų. This is the type of subunit:subunit interaction observed in both DesVI and TylM1, which function as dimers. Strikingly, as can be seen in Figure 5(b), DesVI, TylM1, and KedS8 all differ with respect to the conformation of their N-terminal tails. In DesVI, the N-terminus curls away from the main body of the molecule whereas in TylM1 it projects into the active site such that Tyr 14 lies within hydrogen bonding distance to the dTDP-sugar substrate. In KedS8, the N-terminus adopts ϕ, ψ angles characteristic of a β-strand, and it is this portion of the polypeptide chain that reaches over to another subunit of the tetramer to form the subunit:subunit interface highlighted in Figure 5(c). As a result of this β-strand swapping, a ten-stranded antiparallel β-sheet is formed across the two subunits. The total buried surface area for this interface is 3200 Ų. With respect to amino acid sequence alignments, KedS8 begins to correspond with DesVI at Leu 15 and with TylM1 at Glu 20. Excluding these N-terminal tails, the sequence identities of KedS8 with DesVI and TylM1 are 52% and 43%, respectively.

Figure 5. Figure.

Quaternary structure of KedS8. One of the major subunit:subunit interfaces of the tetramer is shown in (a). This organization is also observed for DesVI and TylM1, which function as dimers. A superposition of the ribbon drawings for DesVI (yellow), TylM1 (purple), and KedS8 (white) is presented in (b). The overall folds are remarkably similar except for the positions of the N-terminal tails. The N-terminal tail in KedS8 is responsible for the change from a dimeric to tetrameric quaternary structure as shown in (c).

The structure of KedS8 described here represents the first model for an N-methyltransferase that functions on a C-4′ sugar amino group. The proposed pathway for l-kedarosamine biosynthesis is based solely on bioinformatics; however, and the activities of the putative enzymes leading up to the formation of the substrate for KedS8 have not been experimentally verified.³ In our hands it has not been possible to produce a dTDP-linked sugar substrate for KedS8 thus far based on the published pathway. Indeed, it is unclear whether KedS8 functions as a mono- or dimethyltransferase. Regardless, the three-dimensional architecture described herein for KedS8 emphasizes that in this particular protein family, the conformations of the N-terminal tails play major roles in the quaternary structures assumed by the proteins, and in some cases the manner in which the active sites are constructed.

Materials and Methods

Cloning of the kedS8 gene

The kedS8 gene was cloned via PCR from Streptoalloteichus sp. ATCC 53650 using Platinum Pfx DNA polymerase (Invitrogen). Primers were designed that incorporated NdeI and XhoI restriction sites. The PCR product was digested with NdeI and XhoI and ligated into pET28T, a laboratory pET28b(+) vector that had been previously modified to incorporate a TEV protease cleavage recognition site after the N-terminal polyhistidine tag.⁸

Protein expression and purification

The pET28t-kedS8 plasmid was utilized to transform Rosetta2(DE3) Escherichia coli cells (Novagen). The cultures were grown in lysogeny broth supplemented with kanamycin and chloramphenicol at 37°C with shaking until an optical density of 0.8 at 600 nm was reached. The flasks were cooled in an ice bath. Protein expression was initiated by the addition of 1 mM isopropyl β-d-1-thiogalactopyranoside. The cultures were allowed to grow at 23°C for 24 h.

The cells were harvested by centrifugation and lysed by sonication on ice. The lysate was cleared by centrifugation, and KedS8 was purified with Ni-NTA resin (Qiagen) according to the manufacturer’s instructions using a lysis/wash buffer of 50 mM sodium phosphate, 300 mM NaCl, and 20 mM imidazole (pH 8.0), and an elution buffer of 50 mM sodium phosphate, 300 mM NaCl, and 250 mM imidazole (pH 8.0). The histidine tag was removed by digestion with TEV protease at a 1:20 protease:KedS8 molar ratio, for 36 h at 4°C. The protease and uncleaved KedS8 were removed by passage over Ni-NTA resin, and the protein was dialyzed against 10 mM Tris-HCl (pH 8.0) and 200 mM NaCl and concentrated to 17.5 mg/ml based on an extinction coefficient of 1.04 (mg/ml)⁻¹ cm⁻¹.

Crystallization and structural analysis

Crystallization conditions were surveyed at both room temperature and 4°C by the hanging drop method of vapor diffusion using a laboratory based sparse matrix screen. In an attempt to produce crystals of KedS8 bound to a dTDP-sugar analog, the initial crystals were grown from precipitant solutions containing 28–33% 2-methyl-2,4-pentanediol, 2.5 mM SAH, 10 mM dTDP-4-deoxy-4-amino-d-quinovose, and 100 mM MOPS (pH 7.0) at 4°C. The crystals belonged to the space group I222 with unit cell dimensions of a=63.3 Å, b=81.0 Å, and c=126.9 Å and one subunit in the asymmetric unit.

Subsequent X-ray data collection using these crystals showed that only SAH had bound in the active site. In light of this result, these crystals were then soaked in a solution containing 35% 2-methyl-2,4-pentanediol, 200 mM NaCl, 2.5 mM SAH, 40 mM dTDP-benzene, and 100 mM MOPS (pH 7.0). Crystals were prepared for data collection at 100 K by transferring to a solution composed of 38% 2-methyl-2,4-pentanediol, 200 mM NaCl, 2.5 mM SAH, 40 mM dTDP-benzene, 3% ethylene glycol, and 100 mM MOPS (pH 7.0).

X-ray data were then collected from these crystals at 100 K with a Bruker AXS Platinum 135 CCD detector controlled by the Proteum software suite (Bruker AXS). The X-ray source was Cu Kα radiation from a Rigaku RU200 X-ray generator equipped with Montel optics and operated at 50 kV and 90 mA. These X-ray data were processed with SAINT version 7.06 A (Bruker AXS) and internally scaled with SADABS version 2005/1 (Bruker AXS). Relevant X-ray data collection statistics are listed in Table1. The KedS8 structure was determined via molecular replacement with the software package PHASER and using as a search model the X-ray coordinates 3BXO from the Protein Data Bank.^4,10 Iterative cycles of model building with COOT and refinement with REFMAC reduced the R_work and R_free to 16.8% and 22.6%, respectively, from 30 to 2.0 Å resolution.^11–13

Table 1.

X-ray data collection statistics and model refinement statistics

	Binary complex
Resolution limits (Å)	30.0–2.0, (2.1–2.0)a
Number of independent reflections	21,803, (2833)
Completeness (%)	97.6, (94.0)
Redundancy	3.4, (2.2)
avg I/avg σ(I)	14.3, (4.2)
Rsym (%)b	4.5, (15.9)
cR-factor (overall)%/no. reflections	17.1/21,803
R-factor (working)%/no. reflections	16.8/20,689
R-factor (free)%/no. reflections	22.6/1114
Number of protein atoms	1907
Number of heteroatoms	221
Average B values
Protein atoms (Å2)	28.2
Ligand (Å2)	28.7
Solvent (Å2)	35.4
Weighted RMS deviations from ideality
Bond lengths (Å)	0.012
Bond angles (°)	1.95
Planar groups (Å)	0.009
Ramachandran regions (%)d
Most favored	93.9
Additionally allowed	6.1
Generously allowed	0.0

^aStatistics for the highest resolution bin.

^bR_sym=(∑|I−|/∑ I) x 100.

^cR-factor=(Σ|F_o−F_c|/Σ|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.⁹

Determination of quaternary structure

The quaternary structures of purified TylM1, known to be a dimer, and KedS8 were analyzed using a Superdex 200 10/300 (GE Healthcare) gel filtration column and an ÄKTA HPLC system. The samples were loaded and run in 10 mM Tris (pH 8.0) and 200 mM NaCl. The column was run at a speed of 0.5 mL/min at ambient temperature. Comparison of retention times with a set of standard proteins from Sigma-Aldrich demonstrated that TylM1 ran as expected for a dimeric protein (molecular weight ∼55,000, retention time 16.8 min) whereas KedS8 ran as expected for a tetrameric species (molecular weight ∼110,000, retention time 14.5 min). Standards used for comparison: alcohol dehydrogenase (retention time 12.5 min, MW=150,000) albumin (retention time 16.3 min, MW=66,000), and carbonic anhydrase (retention time 17.9 min, MW=29,000).

Glossary

dTDP

thymidine diphosphate

HPLC

high performance liquid chromatography

MOPS

3-(N-morpholino)propanesulfonic acid

Ni-NTA

nickel nitrilotriacetic acid

PCR

polymerase chain reaction

TEV

tobacco etch virus

Tris

tris-(hydroxymethyl)aminomethane.