Extremophiles (2015) 19:417–427 DOI 10.1007/s00792-014-0727-9

ORIGINAL PAPER

Characterization of the amino acid residues mediating the unique amino‑sugar‑1‑phosphate acetyltransferase activity of the archaeal ST0452 protein Zilian Zhang · Yasuhiro Shimizu · Yutaka Kawarabayasi 

Received: 12 November 2014 / Accepted: 12 December 2014 / Published online: 8 January 2015 © Springer Japan 2015

Abstract  The ST0452 protein from the thermophilic archaean Sulfolobus tokodaii has been identified as an enzyme with multiple sugar-1-phosphate nucleotidylyltransferase and amino-sugar-1-phosphate acetyltransferase (amino-sugar-1-P AcTase) activities. Analysis of the protein showed that in addition to glucosamine-1-phosphate (GlcN-1-P) AcTase activity, it possesses unique galactosamine-1-phosphate (GalN-1-P) AcTase activity not detected in any other proteins. Comparison of the crystal structures of the ST0452 protein and GlmU from Escherichia coli (EcGlmU), which possesses only GlcN-1-P AcTase activity, showed that the overall sequence identity between these two proteins is less than 25 %, but the amino acid residues predicted to comprise the catalytic center of EcGlmU are conserved in the ST0452 protein. To understand the molecular mechanism by which the ST0452 amino-sugar-1-P AcTase activity recognizes two independent substrates, several ST0452 substitution and truncation mutant proteins were constructed and analyzed. We found that His308 is

essential for both GalN-1-P and GlcN-1-P AcTase activities, whereas Tyr311 and Asn331 are important only for the GalN-1-P AcTase activity. In addition, deletion of the C-terminal 5 or 11 residues showed that the 11-residue C-terminal region exerts a modest stimulatory effect on GalN-1-P AcTase activity but dramatically suppresses GlcN-1-P AcTase activity. This region also appears to make an important contribution to the thermostability of the entire ST0452 protein. Systematic deletions from the C-terminus also demonstrated that the C-terminal region with the β-helix structure has an important role mediating the trimerization of the ST0452 protein. This is the first report of an analysis of a thermostable archaeal enzyme exhibiting multiple amino-sugar-1-P AcTase activities. Keywords  Archaea · Thermophiles · Acetyltransferase · Mutation · Reaction center · Thermophilic enzyme

Introduction Communicated by H. Atomi. Z. Zhang  State Key Laboratory of Marine Environmental Science, Institute of Marine Microbes and Ecospheres, Xiamen University, Xiamen 361005, People’s Republic of China Y. Shimizu · Y. Kawarabayasi (*)  Laboratory of Functional Genomics of Extremophiles. Faculty of Agriculture, Kyushu University, Hakozaki 6‑10‑1 Higashi‑ku, Fukuoka, Fukuoka 812‑8581, Japan e-mail: [email protected] Y. Kawarabayasi  National Institute of Advanced Industrial Science and Technology (AIST), Institute for Health Research, Amagasaki, Hyogo 661‑0974, Japan

Carbohydrate polymers play crucial roles at the bacterial cell surface. Peptidoglycan, for example, forms the bacterial cell wall, which also includes the polysaccharide virulence factors teichoic acid in Gram-positive species and, lipopolysaccharide in the outer membrane of Gram-negative species (Kotnik et al. 2007; Gehring et al. 1996; Raetz 1993). A key component of these carbohydrate polymers is N-acetyl-d-glucosamine (GlcNAc) (Konopka 2012), and UDP-GlcNAc, an activated form of GlcNAc, serves as the substrate for biosynthesis of both teichoic acid and polysaccharides (Leloir 1951). The enzymes involved in the UDP-GlcNAc biosynthetic pathway are thus considered to be potentially effective targets for the development of new antibacterial agents (Kotnik et al. 2007).

13

418

It is well-known that UDP-GlcNAc is produced from fructose-6-phosphate (Frc-6-P) in the four-step reactions, and in this biosynthetic pathway two different types, bacterial and eukaryotic types, has been identified. In the bacterial UDP-GlcNAc biosynthetic pathway, the final two reactions are catalyzed by a bifunctional enzyme, GlmU, which exhibits two independent glucosamine-1-phosphate acetyltransferase (GlcN-1-P AcTase) and GlcNAc-1-phosphate uridyltransferase (GlcNAc-1-P UTase) activities (Gehring et al. 1996). In most archaea, the UDP-GlcNAc biosynthetic pathway was not predicted from their genomic information determined. From the genomic data of the thermophilic archaeon Sulfolobus tokodaii strain 7, the ST0452 protein was detected as a glucose-1-phosphate thymidylyltransferase (Glc-1-P TTase), based on its similarity to bacterial Glc-1-P TTase (Zhang et al. 2005). However, experimental analysis of the enzymatic activity of the ST0452 protein demonstrated that it possessed GlcNAc-1-P nucleotidylyltransferase (GlcNAc-1-P NTase) and N-acetyl-d-galactosamine-1-phosphate uridyltransferase (GalNAc-1-P UTase) activities, as well as the expected Glc-1-P TTase activity (Zhang et al. 2005). Furthermore, surprisingly, experimental analysis of its amino-sugar-1-phosphate AcTase (amino-sugar-1-P AcTase) activity showed that in addition to GlcN-1-P, the AcTase activity of the ST0452 protein accepted galactosamine-1-phosphate (GalN-1-P) as substrate, which has not been seen with any known bacterial GlmU (Zhang et al. 2010). Although the observations summarized above indicates that the catalytic activities of the ST0452 protein toward GlcN-1-P and GlcNAc-1-P are similar to those of Escherichia coli GlmU (EcGlmU) (Gehring et al. 1996), the amino acid sequences of the two proteins share less than 25 % identity. Moreover, only the ST0452 protein exhibits the unique GalN-1-P acetyltransferase (GalN-1-P AcTase) activity, which was not present on EcGlmU (Zhang et al. 2010). Although the difference of activity was clearly detected, how only the ST0452 protein recognized and catalyzed the GalN-1-P as substrate remains unclear. Answering that question could shed light on the mechanism by which only the ST0452 protein recognizes and catalyzes this GalN-1-P substrate as well as on the evolution of this family of proteins. In the present study, therefore, we addressed this issue by examining the effects of introducing several substitution and truncation mutations into the C-terminal domain of the ST0452 protein. The same amino acid residues at the same position in reaction center showed different role in the ST0452 protein and EcGlmU. Also, the C-terminal region indicated an important role for thermostability of the entire ST0452 protein. This is the first experimental analysis of the mechanism of thermostable archaeal multiple amino-sugar-1-P AcTase activities.

13

Extremophiles (2015) 19:417–427

Materials and methods Bacterial strains and chemical reagents The archaeal strain S. tokodaii strain 7 (JCM10545) was obtained from the Japan Collection of Microorganisms (JCM). E. coli strain BL21-Codon Plus(DE3)-RIL, which was used for expression of the recombinant protein, was obtained from Stratagene (La Jolla, CA). GlcN-1-P, GalN1-P, GlcNAc-1-P, UTP, UDP-GlcNAc, acetyl-CoA and the thiol reagent 5,5-dithiobis(2-nitrobenzoic acid) (DTNB) were all purchased from Sigma Chemical Co. (St. Louis, MO). Primers used in this work The nucleotide sequences of all primers and combinations of primers used for PCR amplification of mutant ST0452 genes are listed in Tables 1 and 2, respectively. Comparison of the 3‑D crystal structures Chimera ver. 1.8.1 software was used for comparison of the 3-D crystal structures of the ST0452 protein and EcGlmU. Data for the ST0452 protein (2GGO) and EcGlmU (2OI5) obtained from the PDB database were used in this analysis. Construction of expression vectors encoding site‑directed ST0452 mutant proteins All site-directed mutant ST0452 genes used in this work were constructed using synthetic primers (SP01–SP07) harboring the substituted nucleotide (Table 1). Fragments with mutations replaced the corresponding regions in the parental ST0452H plasmid, which was constructed as described previously (Zhang et al. 2005). Primers P1 and P3 were designed for an earlier study in which an expression vector encoding wild-type ST0452 protein with a hexahistidine tag at its C-terminus was made (Zhang et al. 2005). To substitute His308 with Ala, two overlapping fragments were amplified using the primer pairs P1/SP01 and SP02/ P3. Similarly, to substitute Tyr311 with Ala, two overlapping fragments were amplified using primer pairs P1/SP03 and SP04/P3. After purification, the amplified fragments were digested with the appropriate restriction enzymes (Table  2) and ligated into pET21(b) digested with NdeI and XhoI. To substitute Asn331, Lys337 and Lys340 with Ala, PCR fragments were amplified using primer pairs P1/SP05, SP06/P3 and SP07/P3, respectively. After purification, the amplified fragments were digested with the appropriate restriction enzymes (Table 2), ligated into pST0452H digested with the same restriction enzymes, and confirmed by sequencing. The constructed plasmids were designated pST0452(H308A)H, pST0452(Y311A)

Extremophiles (2015) 19:417–427 Table 1  The primers used in the study

The underlines indicate the recognition site of the restriction enzyme shown in the right column. The bold lower case letters indicate the nucleotides introduced for amino acid substitution

Table 2  Combinations of primers used for construction of ST0452 mutant genes

419 Primer ID

Sequences

Restriction enzyme

P1

ATAGCATATGAAGGCATTTATTCTTGCTGC

NdeI

P3

TCAACTCGAGGACCTTGAAAAACTCACC

XhoI

SP01

TAGCTAAGGgcCGGAATCTTAG

MspI

SP02

TAAGATTCCGgcCCTTAGC

MspI

SP03

AACAgcGCTAAGGTGCGG

HhaI

SP04

ACCTTAGCgcTGTTGGTGATTC

HhaI

SP05

AAACCTTAAGgcAGCTATTAAAGTGCC

SP06 SP07 DP01

TAACTTAAGGTTTGACGAAgcgGAAGT TAACTTAAGGTTTGACGAAAAAGAAGTTgcGGTTAAT TCAACTCGAGACCATAGCCGACATCTCTG

AflII AflII AflII

DP02

TCAACTCGAGGTTTACAACCGCGCCAGGAT

XhoI

DP03

TCAACTCGAGATAAGCACCAATTTTTACGC

XhoI

DP04

TCAACTCGAGGGTAACGTTTATTCCAG

XhoI

DP05

TCAACTCGAGTCCAATAAATGCCCCTAG

XhoI

DP06

TCAACTCGAGCTAACTACTTATTCTTTTTCC

XhoI

DP07

TCAACTCGAGAGCTATTAAAGTGCCTGCACC

XhoI

DP08

TCAACTCGAGATAAGGTCTTAAATAGGAATTTG

XhoI

DP09

TCAACTCGAGCTAATTTTGACTAAATACAAG

XhoI

5′ region Primers

Vectors constructed

3′ region Restriction enzymes

Primers

XhoI

Restriction euzymes

PI

SP01

NdeI/MspI

SP02

P3

MspI/XhoI

pST0452(H308A)H

PI

SP03

NdeI/HhaI

SP04

P3

HhaI/XhoI

pST0452(Y311A)H

PI

SP05

NdeI/AflII SP06

P3

Af1II/XhoI

pST0452(K337A)H

SP07

P3

Af1II/XhoI

pST0452(K340A)H

pST0452(N331A)H

PI

DP01

NdeI/XhoI

pST0452(DC005)H

PI

DP02

NdeI/XhoI

pST0452(DC011)H

PI

DP03

NdeI/XhoI

pST0452(DC021)H

PI

DP04

NdeI/XhoI

pST0452(DC031)H

PI

DP05

NdeI/XhoI

pST0452(DC041)H

PI

DP06

NdeI/XhoI

pST0452(DC051)H

PI

DP07

NdeI/XhoI

pST0452(DC071)H

PI

DP08

NdeI/XhoI

pST0452(DC121)H

PI

DP09

NdeI/XhoI

pST0452(DC171)H

H, pST0452(N331A)H, pST0452(K337A)H and pST0452 (K340A)H. Construction of expression vectors encoding ST0452 C‑terminal deletion mutant proteins Primers DP01–DP09 were designed from the nucleotide sequence of the corresponding regions of ST0452 gene and contained a XhoI site (Table 1). To construct expression vectors encoding a series of ST0452 C-terminal deletion mutants

with hexahistidine tags at their C-termini, nine primer pairs were used. The temperature profiles used for PCR amplification were the same as described previously (Zhang et al. 2005). After purification, the amplified fragments were digested with NdeI and XhoI, inserted into pET21(b) digested with the same restriction enzymes, and confirmed by sequencing. The constructed plasmids harboring the respective PCR fragments for expression of ST0452 deletion mutant proteins with hexahistidine tags at their C-termini were designated pST0452(DC005)H, pST0452(DC011)H,

13

420

Extremophiles (2015) 19:417–427

pST0452(DC021)H, pST0452(DC031)H, pST0452(DC041) H, pST0452(DC051)H, pST0452(DC071)H, pST0452 (DC121)H and pST0452(DC171)H.

HPLC and measured based on the absorbance at 254 nm (A254), as described previously (Zhang et al. 2005).

Expression and purification of the recombinant proteins

Results

Wild-type ST0452 protein, the ST0452 substitution mutant proteins and the ST0452 C-terminal deletion mutant proteins were all expressed in E. coli strain BL21-Codon Plus(DE3)-RIL and purified as described previously (Zhang et al. 2005). The purified ST0452 proteins were then stored at 4 °C. The protein concentrations were determined using a BCA assay kit (Pierce Biotechnology, Rockford, IL).

Comparison of the crystal structures of the ST0452 protein and EcGlmU

Assay of amino‑sugar‑1‑P acetyltransferase activity Amino-sugar-1-phosphate acetyltransferase activity was measured as the amount of CoA produced from acetylCoA using the previously described method (Zhang et al. 2010). The reaction was run in a 10-μL reaction mixture containing 50 mM Tris–HCl (pH 7.5), 2 mM MgCl2, 2 mM acetyl-CoA and 2 mM GlcN-1-P or GalN-1-P. After incubating this reaction mixture for 1 min at 80 °C, the reaction was started by adding 50 ng of recombinant wild-type or mutant ST0452 proteins. The reaction then proceeded for appropriate period at 80 °C. To stop the reaction, 40 μL of solution containing 50 mM Tris–HCl (pH 7.5) and 6.4 M guanidine hydrochloride were added. Thereafter, 50 μL of solution containing 50 mM Tris–HCl (pH 7.5), 1 mM EDTA and 0.5 mM DTNB solution were added, and the absorbance at 412 nm (A412) was measured (Riddles et al. 1983). The amounts of CoA released were calculated from a standard curve using 10,364 M−1 cm−1 as the molar absorption coefficient for the CoA molecule. To determine the kinetic parameters, the concentration of either acetylCoA or GlcN-1-P/GalN-1-P was varied while keeping the concentration of the other substrate 5–10 times higher than its Km value. Double-reciprocal plots of the initial velocities were prepared to determine Km and kcat. Assay of GlcNAc‑1‑P UTase activity GlcNAc-1-P UTase activity was analyzed based on the amount of UDP-GlcNAc produced in the reaction mixture. The assay for the forward reaction, production of the UDPGlcNAc from GlcNAc-1-P and UTP, was performed in a 30-μL reaction mixture containing 50 mM Tris–HCl (pH 7.5), 2 mM MgCl2, 10 mM GlcNAc-1-P and 0.1 mM UTP. After incubating this reaction mixture for 1 min at 80 °C, the reaction was started by adding 50 ng of purified recombinant protein. After incubation at 80 °C for 1 min, 300 μl of 500 mM KH2PO4 were added to stop the reaction. UDPGlcNAc, the product of the reaction, was then separated by

13

To assess the similarity between the 3-D structures of the ST0452 protein and EcGlmU, we obtained their crystal structure data from the PDB, which showed that the overall structures of these two proteins differed greatly (Fig. 1a). In both, however, left-handed β-helixes (LβHs) were detected within the C-terminal domain, which prompted us to separately compare the structures of the proteins’ N- and C-terminal domains. The structures from the N-terminus to residue 210 of the ST0452 protein and 226 of EcGlmU are shown in Fig. 1b, while the remainder of the proteins, from residue 211 of the ST0452 protein and 227 of EcGlmU to the C-terminus, are shown in Fig. 1c. Although the overall structures show less than 25 % amino acid sequence identity, the respective N- and C-terminal domains do show structural similarity, with root-mean-square deviation (RMSD) of 1.22 and 0.617 Å, respectively. And because they contain the catalytic centers for the acetyltransferase activity, the C-terminal domains were compared in more detail. The shapes of the ST0452 protein and EcGlmU C-terminal domains appeared similar, though three structural differences were identified (Fig. 1c). First, the C-terminal LβH contains eight turns in the ST0452 protein but nine turns in EcGlmU. Second, the extra loop region in the ST0452 protein is 16 residues longer than in EcGlmU (Fig. 2). Nonetheless, the structural positions at the start and end of the ST0452 extra loop nearly match those of the EcGlmU extra loop (Figs. 1c, 2). Finally, the C-terminal tail region of the ST0452 protein (two amino acid residues long) is 17 residues shorter than that of EcGlmU (19 residues long) (Fig. 2). From the 3-D crystal structure of EcGlmU in complex with its substrate for GlcN-1-P AcTase, it was demonstrated that within the GlcN-1-P AcTase catalytic center of EcGlmU, four residues, His363, Tyr366, Asn386 and Lys392, were identified as amino acid residues interacted with its substrates (Olsen and Roderick 2001; Olsen et al. 2007). Despite the low amino acid sequence identity between EcGlmU and the ST0452 protein, those four amino acid residues are conserved in the ST0452 protein as His308, Tyr311, Asn331 and Lys337 (Fig. 2). Analysis of the ST0452 substitution mutant proteins Despite the conservation of the amino acid residues predicted to be important for the GlcN-1-P AcTase activity

Extremophiles (2015) 19:417–427

421

Fig. 1  3-D structures of EcGlmU and the ST0452 protein. a Overall crystal structures of EcGlmU (PDB: 2OI5) (left) and the ST0452 protein (PDB: 2GGO) (right). b, c Structural comparison of the N-terminal (b) and C-terminal (c) domains of EcGlmU (brown) and the ST0452 protein (blue). These domains were compared using Chimera 1.8.1. Views from different angles are shown on the left and right

of EcGlmU, the ST0452 protein can catalyze both GlcN1-P and GalN-1-P as substrates, even though EcGlmU can catalyze only GlcN-1-P. To determine which, if any, of the residues conserved in the ST0452 protein recognize GalN1-P as a substrate, Ala was substituted for His308 (H308A), Tyr311 (Y311A), Asn331 (N331A) or Lys337 (K337A). In addition, because Lys340 is situated in close proximity to Lys337, it too was substituted with Ala (K340A). The H308A mutant showed only 7.8 % of the GalN-1-P AcTase activity and 0.7 % of the GlcN-1-P AcTase activity

of wild-type ST0452 protein (Table 3), revealing that this His residue is essential for both amino-sugar-1-P AcTase activities of the ST0452 protein. By contrast, whereas the Y311A and N331A mutants respectively showed only 3.3 and 3.1 % of the GalN-1-P AcTase activity of wild-type ST0452 protein, they respectively, exhibited 118.4 and 46.1 % of the GlcN-1-P AcTase activity of the wild-type ST0452 protein. This means that both Tyr311 and Asn331 are essential for GalN-1-P AcTase activity of the ST0452 protein, but that Tyr311 slightly inhibits and Asn331

13

422

Extremophiles (2015) 19:417–427

A

B

C

Fig. 2  Amino acid sequence and alignment of the C-terminal regions of the ST0452 protein and EcGlmU. a, b Amino acid residues are aligned for each turn of the LβH structure in the C-terminal domains of the ST0452 protein (residue 234 to the C-terminus; a) and EcGlmU (residue 260 to the C-terminus; b). The residues in parentheses and brackets comprise the extra loops and C-terminal tail regions, respectively. The underlined residues are within sheet strucTable 3  Enzymatic activities of the ST0452 substitution mutants

Protein

tures. c Alignment of the acetyltransferase domains of the ST0452 protein and EcGlmU (EcGlmU, YP_491699). The underlined residues are within sheet structures. Residues shown in bold with a square were chosen for mutation. The vertical lines indicate the positions of the C-terminus after systematic deletion in the C-terminal domain

GlcN-1-P acetyltransferase

GalN-l-P acetyltransferase

Specific activity (μmol/min/ Relative activity (%) Specific activity (μmol/min/ Relative mg protein) mg protien) activity (%)

The relative activities are shown as percentages of the activity of wild-type ST0452 protein. All experiments were repeated for three times

ST0452 H308A Y311A N331A K377A

50.0 ± 0.6 0.35 ± 0.08 59.2 ± 3.6 23.1 ± 0.2 68.9 ± 2.3

0.7 118.4 46.1 137.7

40.0 ± 1.1 3.0 ± 1.5 1.31 ± 0.25 1.23 ± 0.10 33.0 ± 1.1

7.7 3.3 3.1 82.6

K340A

73.5 ± 2.5

147.1

25.3 ± 1.3

63.3

slightly enhances GlcN-1-P AcTase activity of the ST0452 protein. The functions of these amino acid residues clearly differed from the corresponding residues in EcGlmU. In addition, substitution of Lys337 and Lys340, respectively reduced GalN-1-P AcTase activity to 82.6 and 63.3 % of that of the wild-type ST0452 protein, but increased GlcN1-P AcTase activity to 137.7 and 147.1 % of that of the wild-type ST0452 protein (Table 3). This indicates that these two residues, Lys337 and Lys340 in the extra loop region, act to enhance the GalN-1-P AcTase activity but

13

suppress the GlcN-1-P AcTase activity of the ST0452 protein. Analysis of the ST0452 C‑terminal deletion mutant proteins The length of the C-terminal tail regions of EcGlmU and the GlmU from Mycobacterium tuberculosis (MtGlmU) are 19 and 45 amino acid residues long, respectively. These tail regions are situated at the bottom of the protein’s trimeric structure and

Extremophiles (2015) 19:417–427

423

Table 4  Enzymatic activities of the ST0452 C-terminal deletion mutants Protein

GlcNAc-1-P UTase (μmol/min/mg protein) GlcN-l-P acetyltransferase (μmol/min/mg protein)

ST0452 2.2 ± 0.1 DC005 2.0 ± 0.1 DC011

2.6 ± 0.1

GalN-l-P acetyltransferase (μmol/min/mg protein)

67.4 ± 1.2 323.5 ± 13.2

47.6 ± 0.7 38.0 ± 1.5

1,130.7 ± 45.6

29.6 ± 0.6

All experiments were repeated for three times

Table 5  Kinetic properties of the acetyltransferase activity of wild-type and C-terminal deletion forms of the ST0452 protein Protein

Substrates

Km (mM)

ST0452

Acetyl-CoA GlcN-l-P GalN-l-P Acetyl-CoA

0.63 ± 0.02 0.59 ± 0.05 1.71 ± 0.2 0.52 ± 0.00

GlcN-l-P GalN-l-P Acetyl-CoA GlcN-l-P

0.56 ± 0.04 0.84 ± 0.05 1.55 ± 0.01 1.69 ± 0.07

GalN-l-P

0.66 ± 0.09

DC005

DC011

kcat (s−1)

kcat/Km (mM−1s−1)

123.2 ± 4.6 69.7 ± 6.9

211.0 ± 14.1 40.8 ± 3.7

490.6 ± 35.6 17.9 ± 0.3

869.2 ± 2.1 21.5 ± 1.0

2310.8 ± 12.4

1488.9 ± 13.4

12.6 ± 1.3

19.6 ± 0.5

All experiments were repeated for three times

enter the region bordering the other two subunits (Olsen and Roderick 2001; Olsen et al. 2007; Jagtap et al. 2012). In addition, Trp residues located within this tail region play an essential role for the GlcN-1-P AcTase activity of MtGlmU (Jagtap et al. 2012). In the ST0452 protein, by contrast, the C-terminal tail region is only two amino acid residues long (Fig. 2), and there is no Trp residue within this region. Despite these differences in the structures of the tail regions among the two bacterial GlmUs and the ST0452 protein, all three proteins were able to acetylate GlcN-1-P; only the ST0452 protein catalyzed the GalN-1-P AcTase reaction. This suggested that the C-terminal tail region of the ST0452 protein might be important for recognition of the multiple substrates for amino-sugar-1-P AcTase activity. To test that idea, we expressed two deletion mutants, DC005 and DC011, in E. coli (Fig. 2). After purification of the proteins using nickel affinity column chromatography, the GalN-1-P AcTase, GlcN-1-P AcTase and GlcNAc-1-P UTase activities were measured at 80 °C for 1 min, which is shorter than the denaturing period of DC011. As shown in Table 4, the GlcNAc-1-P UTase activity of DC005 was reduced by 10 %, as compared to that of the wild-type ST0452 protein, while the activity of DC011 was increased by 18 %, indicating that the enzymatic activity in the N-terminal region was not greatly affected by truncation of the C-terminal 5 or 11 residues. As the amino-sugar-1-P AcTase activities, DC005 and DC011, respectively showed 20 and 38 % less GalN-1-P AcTase activity than that of the wild-type ST0452 protein,

revealing that this region has enhance effect on the GalN1-P AcTase activity of the ST0452 protein. Conversely, the GlcN-1-P AcTase activities of DC005 and DC011 were respectively, 4.8 and 16.8 times higher than that of the wild-type ST0452 protein. Thus, the C-terminal 11-residue region of the ST0452 protein appears to exert an enhance effect on its GalN-1-P AcTase activity and a strongly inhibitory effect on its GlcN-1-P AcTase activity. This feature is markedly different from the essential role played by the C-terminal tail region in MtGlmU GlcN-1-P AcTase activity (Jagtap et al. 2012). To better understand the dramatic increase in GlcN-1-P AcTase activity caused by deletion of the C-terminal 5 or 11 residues of the ST0452 protein, the kinetic constants for GlcN-1-P and GalN-1-P were determined. Because DC011 protein showed little stability at 80 °C, assays were carried out for 1 min at 80 °C. Table 5 shows the apparent Km values for acetyl-CoA and GlcN-1-P with DC005 or DC011 protein. Both Km values for the GlcN-1-P AcTase activity of DC005 were approximately 10 % lower than that of the wild-type ST0452 protein. On the other hand, both Km values for the GlcN-1-P AcTase activity of DC011 were 2.6 times higher than that of the wild-type ST0452 protein. These observations indicate that the C-terminal 5 residues exert an inhibitory effect on the affinity of the acetyl-CoA and GlcN-1-P substrates, and that the C-terminal 11 residues play an important role in the binding of GlcN-1-P and acetyl-CoA. By contrast, the kcat values for DC005 and

13

424

DC011 were respectively about 4 times and 18.8 times higher than for wild-type ST0452 protein (Table 5), indicating the C-terminal 11 amino acids suppressed the rate of the GlcN1-P AcTase catalyzed reaction. Analysis of the GalN-1-P AcTase activities of DC005 and DC011 showed the Km values to be 51 and 61 % lower than the value for the wild-type ST0452 protein, which means the truncated mutants are able to utilize substrates at lower concentrations. The kcat values for GalN-1-P with DC005 and DC011 were only 26 and 18 % of that with the wild-type ST0452 protein. These results indicate that the C-terminal residues of the ST0452 protein enhance the turnover rate of its GalN-1-P AcTase catalytic activity and slightly suppress substrate binding. Taken together, these observations show that the same 11-residue C-terminal region exerts opposite effects with two similar but independent substrates, GlcN-1-P and GalN-1-P. We next investigated the effect of the C-terminal 11-residue region on the thermostability of the ST0452 protein. When DC005 and DC011 were incubated for 5 min at temperatures ranging from 37 °C to 80 °C, DC005 showed the same thermostability as wild-type ST0452 protein, whereas DC011 denatured and became insoluble form by 5-min treatment at 80 °C (Fig. 3a). In fact, DC011 began to denature by 2 min at 80 °C, but most of the protein remained in solution for the first 1 min at 80 °C (Fig. 3b). These observations indicate the C-terminal 11 amino acids of the ST0452 protein are also important for the thermostability of the entire protein at 80 °C. Role of the C‑terminal domain in stabilizing the ST0452 protein It was previously shown that the ST0452 truncation mutant protein lacking the entire 170 residues of the C-terminal domain became completely insoluble after 5 min treatment at 70 °C (Zhang et al. 2005). The present observations that deletion of the C-terminal 11 amino acids of the ST0452 protein also significantly affected the thermostability of the entire protein, prompted us to test the effects on thermostability of deleting larger C-terminal portions (Fig. 4). Among the deletion mutants constructed, only DC051 and DC171 could be expressed in a soluble form. The remaining proteins, DC021, DC031, DC041, DC071 and DC121, were produced in an insoluble form or aggregated immediately after purification. As a result, their stability could not be analyzed. Examination of their thermostability showed that DC051 became completely insoluble after 5 min treatment at 60 °C, while DC171 was insoluble after 5 min treatment at 70 °C. Bacterial GlmU reportedly assumes a trimeric structure (Olsen and Roderick 2001; Olsen et al. 2007). When the tertiary structures of the deletion ST0452 proteins were analyzed using gel filtration, we found that DC005 and

13

Extremophiles (2015) 19:417–427

A

C 37 40 50 60 70 80 ST0452 DC005 DC011

B

C

1

2

3

4

5 min

Fig. 3  Effect of C-terminal deletion on the thermostability of the ST0452 protein. a SDS-PAGE of wild-type ST0452 protein and deletion mutants lacking the 5 (DC005) or 11 (DC011) C-terminal residues after incubation for 5 min at the indicated temperature. b SDSPAGE of DC011 after incubation at 80 °C for the indicating times. The protein concentration used for the heat treatment was 1 mg/ml in buffer containing 50 mM Tris–HCl (pH 7.5) and 2 mM MgCl2. After heating, the solution was centrifuged at 20,000×g for 10 min at 4 °C. The protein remaining in the soluble fraction was used for SDSPAGE analysis. Following electrophoresis, the proteins were stained with Coomassie Brilliant Blue R-250. Lane C contains the proteins shown in the same panel without heat treatment

DC011 proteins retained the same tertiary structures as the wild-type ST0452 protein (data not shown). By contrast, other two mutant proteins, DC051 and DC171, assumed monomeric structures. This suggests formation of the trimeric structure is necessary for stability at high temperature (>70 °C) but not for solubility. Thus, the C-terminal domain of the ST0452 protein, with its LβH structure, appears to be essential for the formation of its trimeric form and, in turn, the high stability of the entire ST0452 protein.

Discussion The ST0452 protein from the acidothermophilic archaeon S. tokodaii was originally predicted to be a Glc-1-P TTase, based on its sequence similarity to bacterial enzymes. However, determination of the actual activities of the protein showed that the ST0452 protein was capable of catalyzing GalN-1-P AcTase and GalNAc-1-P UTase reactions as well as the last two reactions in the UDP-GlcNAc biosynthetic pathway (Zhang et al. 2010). The presence of two amino-sugar-1-P AcTase activities, GalN-1-P AcTase and GlcN-1-P AcTase activities, distinguishes the ST0452 protein from known related bacterial GlmU enzymes. Nonetheless, the ability of the ST0452 protein to catalyze the last two reactions in the bacteria-type UDP-GlcNAc biosynthetic pathway shows a functional similarity to bacterial GlmU enzymes, despite sharing less than 25 % sequence

Extremophiles (2015) 19:417–427

A

425

210 aa

191 aa

B

C 37 40 50 60 70 80

ST0452 DC005 DC011 DC021

NS

DC031

NS

DC041

NS

DC051 DC071

NS

DC121

NS

DC171 Fig. 4  Structures and thermostability of the ST0452 protein and its C-terminal deletion mutants. a Schematic structures of wildtype ST0452 protein and the indicated C-terminal deletion mutants. Black and open bars indicate the N-terminal nucleotidylyltransferase domain (residues 1–210) and the C-terminal acetyltransferase domain (residues 211–401), respectively. b SDS-PAGE of the corresponding proteins on the left after incubation for 5 min at the indicated temperatures. Following electrophoresis, the proteins were stained with

Coomassie Brilliant Blue R-250. NS indicates expression of only insoluble protein. The protein concentration used for heat treatment was 1 mg/ml in buffer containing 50 mM Tris–HCl (pH 7.5) and 2 mM MgCl2. After heat treatment, the solution was centrifuged at 20,000×g for 10 min at 4 °C. The protein remaining in the soluble fraction was used for SDS-PAGE analysis. Lane C contains the proteins shown in the same panels without heat treatment

identity with EcGlmU and MtGlmU. In addition, although the overall 3-D crystal structure of the ST0452 protein (PDB: 2GGO) differs from that of EcGlmU (PDB: 2OI5) (Fig.  1a), separate comparison of their N- and C-terminal domains revealed significant 3-D structural similarities (Fig.  1b, c), with conservation of the amino acid residues in the predicted catalytic center of the amino-sugar-1-P AcTase reaction. That said there is a clear difference in the amino-sugar-1-P AcTase activities of the ST0452 protein and EcGlmU: the ST0452 protein can catalyze the acetylation of both GlcN-1-P and GalN-1-P but bacterial GlmU can catalyze acetylation of only GlcN-1-P. To explore the mechanism underlying the ST0452 protein’s multiple amino-sugar-1-P AcTase activities, detailed analysis of the ST0452 mutant proteins containing substitution at amino acid residues within the predicted catalytic center or truncation of the C-terminal domain enabled us to identify the residues and domains essential for or regulating the amino-sugar-1-P AcTase activities of the ST0452 protein. Substituting Ala for His308, corresponding to His363 in EcGlmU and His374 in MtGlmU, diminished both

amino-sugar-1-P AcTase activities of the ST0452 protein. This observation is consistent with the previously obtained effect of substituting His374 in MtGlmU (Jagtap et al. 2012), and indicates that His308 is essential and key residue for the amino-sugar-1-P AcTase activity of the ST0452 protein and similar proteins. Substituting Ala for Asn331, corresponding to Asn386 in EcGlmU and Asn397 in MtGlmU, strikingly diminished the GalN-1-P AcTase activity of the ST0452 protein, but the GlcN-1-P AcTase activity was diminished by only about half. This result differs from the finding that substituting Ala for the equivalent residue (Asn397) in MtGlmU diminished GlcN-1-P AcTase activity by nearly 95 %, indicating Asn397 to be essential for MtGlmU’s GlcN-1-P AcTase activity. By contrast, whereas Asn331 in the ST0452 protein is essential for the GalN-1-P AcTase activity, it is much less important but not essential for the GlcN-1-P AcTase activity. Thus, although this Asn residue is present at the same structural location in MtGlmU and the ST0452 protein, its function differs. This suggests the amino acid residues surrounding Asn331 are also important

13

426

determinants of substrate specificity, but details of the mechanism by which substrate recognition is determined will require further study. A Trp residue was previously detected within the C-terminal tail region of both MtGlmU and EcGlmU, and was shown as an essential for GlcN-1-P AcTase activity by substitution experiment on MtGlmU (Jagtap et al. 2012). This Trp residue is not present in the ST0452 protein, which has a C-terminal tail region of only two residues, and is thus not required for either of the amino-sugar-1-P AcTase activities of the ST0452 protein. It can therefore be concluded that the reaction mechanism for the amino-sugar-1-P AcTase activity used by the ST0452 protein differs from that used by bacterial GlmU. We also found that substituting Ala for Tyr311, Lys337 or Lys340 respectively reduced the GalN-1-P AcTase activity of the ST0452 protein to 3.3, 82 or 63 % of the wildtype ST0452 protein. By contrast, these three substitutions enhanced of the ST0452 GlcN-1-P AcTase activity to 118, 138 or 147 % of the wild-type ST0452 activity. These three amino acid residues thus act to enhance GalN-1-P AcTase activity but suppress the GlcN-1-P AcTase activity of the ST0452 protein. Two amino acid residues, Tyr311 and Asn331, were found to be essential for GalN-1-P AcTase activity, but not for GlcN-1-P AcTase activity. These two residues were not directly interacted with C4 OH group of GlcN-1-P or GalN-1-P molecules. However, difference of localization or angle of these two substrates within the catalytic center of the amino-sugar-1-P AcTase activity on the ST0452 protein might be thought as reason for different effect on two amino-sugar-1-P AcTase activities by the same substitution in the ST0452 protein. To confirm this hypothesis, co-crystal structures of the wild-type and these two mutant ST0452 proteins with these two substrates should be determined. Deletion analysis of the C-terminal tail region of the ST0452 protein showed that removaling the C-terminal 5 or 11 residues reduced GalN-1-P AcTase activity to 30 or 38 % of the wild-type activity, respectively, but increased GlcN-1-P AcTase activity 4.8 or 16.8 times. These results clearly differ from the dramatic reduction of GlcN-1-P AcTase activity introduced by deletion of the C-terminal 37 residues from MtGlmU. It appears that the C-terminal 11 residues of the ST0452 protein are somewhat important for GalN-1-P AcTase activity, but profoundly suppress the GlcN-1-P AcTase activity. The mechanism underlying the different effects on the similar but independent substrates remains unclear at present. All substitution and truncation mutants of the ST0452 protein except for H308A and N331A indicated decreasing the GalN-1-P AcTase activity and increasing the GlcN-1-P AcTase activity. These observations revealed that Tyr311, Lys337 and Lys340 plus C-terminal 11-residue region of

13

Extremophiles (2015) 19:417–427

the ST0452 protein enhance its GalN-1-P AcTase activity and suppress its GlcN-1-P AcTase activity. It can be thought that this role was lost in bacterial similar enzymes, thus, GalN-1-P AcTase activity was not detected on these enzymes. Therefore, only incorporation or utilization pathway of GalNAc was detected in E. coli (Bricnkkötter et al. 2000). In contrast, reliable amounts of UDP-GalNAc and UDP-GlcNAc were detected in a thermophilic archaeon, Pyrococcus furiosus, by NMR spectroscopy (Ramakrishnan et al. 1997). It hypothesizes that the archaeal enzyme’s high GalN-1-P AcTase activity, which is not detected on the bacterial and eukaryotic similar enzymes, supports the production of the UDP-GalNAc in archaeal cells. Despite the structural similarities between the N- and C-termini of the ST0452 protein and those of EcGlmU, the thermostabilities of the two proteins differ greatly, as EcGlmU is a mesophilic enzyme. Thus, the structures of these proteins do not correlate directly with their thermostability. In the ST0452 protein, the percent Glu and Lys contents are 8.0 and 10.0 %, respectively, which is higher than in EcGlmU (5.7 and 5.5 %, respectively). We would therefore predict that the additional ion pairs formed by the Glu and Lys residues help to stabilize the correct fold of the ST0452 protein at high temperatures, though this remains to be tested. Given that the ST0452 protein contains only two Cys residues, it is unlikely that Cys–Cys bonds contribute to its thermostability. The directions of the hinge regions in the C-terminal domains of the ST0452 protein and EcGlmU differ by approximately, 90 degrees. Consequently, the N-terminal domain is situated on top of the C-terminal domain in EcGlmU, while it is at the side of the C-terminal domain in the ST0452 protein. With those configurations, the distance between the GlcN-1-P AcTase and GlcNAc-1-P UTase catalytic centers is smaller in the ST0452 protein than the mesophilic bacterial GlmU. It might be expected that this situation would be convenient for carrying out contiguous reactions within a single enzyme, but to better understand the meaning of the difference of the positions of the two domains between these two proteins, more detailed analysis of the progression of the reactions will be required. Although the mechanism for the increase in activity induced by some substitutions and truncations is still unknown, this effect may be a useful feature that can be exploited for commercial application of this enzyme. Our group previously reported increasing the GlcNAc-1-P UTase activity of the same ST0452 protein (Zhang et al. 2007). It may be that an ST0452 variant protein exhibiting enhanced GlcN-1-P AcTase and GlcNAc-1-P UTase activities can be created by introducing two or more mutations into the ST0452 protein. In sum, although the details of the mechanisms underlying most of the features of the ST0452 protein remain

Extremophiles (2015) 19:417–427

unclear, this first report on an archaeal thermostable amino-sugar-1-P AcTase nonetheless provides valuable information about the mechanism of substrate recognition by an archaeal amino-sugar-1-P AcTase, as well as its thermostability. Acknowledgments  This work was supported in part by a grant-inaid for scientific research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. It was also supported in part by the Institute for Fermentation, Osaka (IFO).

References Bricnkkötter A, Klöß H, Alpert C-A, Lengeler JW (2000) Pathways for the utilization of N-acetyl-galactosamine and galactosamine in Escherichia coli. Mol Microbiol 37:125–135 Gehring AM, Lees WJ, Mindiola DJ, Walsh CT, Brown ED (1996) Acetyltransfer precedes uridylyltransfer in the formation of UDPN-acetylglucosamine in separable active sites of the bifunctional GlmU protein of Escherichia coli. Biochemistry 35:579–585 Jagtap PK, Soni V, Vithani N, Jhingan GD, Bais VS, Nandicoori VK, Prakash B (2012) Substrate-bound crystal structures reveal features unique to Mycobacterium tuberculosis N-acetyl-glucosamine 1-phosphate uridyltransferase and a catalytic mechanism for acetyl transfer. J Biol Chem 287:39524–39537 Konopka JB (2012) N-acetylglucosamine functions in cell signaling. Scientifica (Cairo). doi:10.6064/2012/489208 Kotnik M, Anderluh PS, Prezelj A (2007) Development of novel inhibitors targeting intracellular steps of peptidoglycan biosynthesis. Curr Pharm Des 13:2283–2309

427 Leloir LF (1951) The enzymatic transformation of uridine diphosphate glucose into a galactose derivative. Arch Biochem Biophys 33:186–190 Olsen LR, Roderick SL (2001) Structure of the Escherichia coli GlmU pyrophosphorylase and acetyltransferase active sites. Biochemistry 40:1913–1921 Olsen LR, Vetting MW, Roderick SL (2007) Structure of the E. coli bifunctional GlmU acetyltransferase active site with substrates and products. Protein Sci 16:1230–1235 Raetz C (1993) Bacterial endotoxins: extraordinary lipids that activate eucaryotic signal transduction. J Bacteriol 175:5745 Ramakrishnan V, Teng Q, Adams MWW (1997) Characterization of UDP amino sugars as major phosphocompounds in the hyperthermophilic archaeon Pyrococcus furiosus. J Bacteriol 179:1505–1512 Riddles PW, Blakeley RL, Zerner B (1983) Reassessment of Ellman’s reagent. Methods Enzymol 91:49–60 Zhang Z, Tsujimura M, Akutsu J, Sasaki M, Tajima H, Kawarabayasi Y (2005) Identification of an extremely thermostable enzyme with dual sugar-1-phosphate nucleotidylyltransferase activities from an acidothermophilic archaeon, Sulfolobus tokodaii strain 7. J Biol Chem 280:9698–9705 Zhang Z, Akutsu J, Tsujimura M, Kawarabayasi Y (2007) Increasing in Archaeal GlcNAc-1-P uridyltransferase activity by targeted mutagenesis while retaining its extreme thermostability. J Biochem 141:553–562 Zhang Z, Akutsu J, Kawarabayasi Y (2010) Identification of novel acetyltransferase activity on the thermostable protein ST0452 from Sulfolobus tokodaii strain 7. J Bacteriol 192:3287–3293

13