N1-Aminopropylagmatine, a New Polyamine Produced as a Key Intermediate in Polyamine Biosynthesis of an Extreme Thermophile, Thermus thermophilus *

In the extreme thermophile Thermus thermophilus, a disruption mutant of a gene homologous to speB (coding for agmatinase = agmatine ureohydrolase) accumulated N1-aminopropylagmatine (N8-amidino-1,8-diamino-4-azaoctane, N8-amidinospermidine), a new compound, whereas all other polyamines produced by the wild-type strain were absent from the cells. Double disruption of speB and speE (polyamine aminopropyltransferase) resulted in the disappearance of N1-aminopropylagmatine and the accumulation of agmatine. These results suggested the following. 1) N1-Aminopropylagmatine is produced from agmatine by the action of an enzyme coded by speE. 2) N1-Aminopropylagmatine is a metabolic intermediate in the biosynthesis of unique polyamines found in the thermophile. 3) N1-Aminopropylagmatine is a substrate of the SpeB homolog. They further suggest a new biosynthetic pathway in T. thermophilus, by which polyamines are formed from agmatine via N1-aminopropylagmatine. To confirm our speculation, we purified the expression product of the speB homolog and confirmed that the enzyme hydrolyzes N1-aminopropylagmatine to spermidine but does not act on agmatine.

Polyamines play important roles in cell proliferation and cell differentiation. Common polyamines such as putrescine, spermidine, and spermine are distributed ubiquitously in cells and tissues at relatively high concentrations (1,2).
Thermus thermophilus, of which the genome project was completed using two strains, HB8 and HB27 (Structural-Biological Whole Cell Project at www.srg.harima.riken.go.jp/ thermus/j_index.htm and see Ref. 3, respectively), produces a variety of polyamines including unusually long polyamines and branched ones (4) (see Fig. 9C). These long and branched polyamines have a marked effect of protecting and stabilizing nu-cleic acids (5,6) and of activating cell-free polypeptide synthesis at high temperature (7)(8)(9).
In many organisms, such as bacteria, yeast, animals, and plants, the first step of polyamine biosynthesis is production of putrescine by decarboxylation of L-ornithine (see Fig. 9A). An additional or alternative pathway of putrescine biosynthesis that is often seen in plants and sometimes in bacteria is decarboxylation of L-arginine followed by hydrolysis of agmatine. Agmatine ureohydrolase or agmatinase, coded by the speB gene, catalyzes this second reaction. The next step is production of spermidine and spermine by the addition of an aminopropyl group to putrescine and spermidine, respectively. This reaction is catalyzed by spermidine or spermine synthase (putrescine/spermidine aminopropyltransferase) coded by the speE gene (1).
To investigate the polyamine biosynthetic pathway in T. thermophilus, we constructed a disruption strain of the speB gene homolog of T. thermophilus. Disruption of the speB gene homolog resulted in drastic reduction of triamines, longer and branched polyamines without accumulation of agmatine, and in accumulation of an unknown compound. Double disruption of speB and speE gene homologs resulted in disappearance of this compound and accumulation of agmatine in the cells. The new compound was identified as N 1 -aminopropylagmatine (N 8amidino-1,8-diamino-4-azaoctane, N 8 -amidinospermidine) by comparison with the chemically synthesized authentic compound. In vitro reactions revealed that SpeE is responsible for the production of N 1 -aminopropylagmatine, and SpeB converts N 1 -aminopropylagmatine to spermidine.

Strains and Culture Conditions
The strains of T. thermophilus and Escherichia coli and plasmids used in this study are listed in Table I. The rich growth and minimum media for T. thermophilus were as described previously (10). Leucine, isoleucine, and uracil (50 g/ml each) were included in the minimum medium. Media were solidified as described (11).

Construction of Plasmids
All nucleotide sequences of T. thermophilus HB8 used in this study were kindly provided by Dr. Seiki Kuramitsu of Osaka University. Construction of pSBKm and pSEPE is shown in Fig. 1. PCR was carried out for 25 cycles (94°C for 0.5 min, 55°C for 0.5 min, and 72°C for 1 min) with LA Taq in GC buffer (Takara Bio) using pairs of oligonucleotide primers listed in Table II. Two pairs of primers, speBKpnUp5Ј-speBHinUp3Ј and speBEcoDw5Ј-speBXbaDw3Ј, were used to construct pSBKm; and an additional two pairs, speEup5Јkpn-speEup3Јhin and speEdw5Јxba-speEdw3Јsac, were used to construct pESE. Genomic DNA of T. thermophilus HB8 was used as the PCR template. PCR products were digested with restriction endonucleases listed in Table II and cloned into pBluescript SK ϩ or pBHTK (14) to construct pSBUKm, pSBD, pSEUKm, and pSED. The 3Ј-half of speB and the downstream * This work was supported in part by Grants-in-aid for Scientific Research 11794038 and 1665704 and by grants-in-aid for promoting bioventures in private universities from the Ministry of Education, Culture, Sports, Science and Technology, the Japanese Government. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The region of speE were inserted into pSBUKm and pSEUKm to prepare pSBKm and pSEKm, respectively. The HTK region of pSEKm was replaced with the pyrE gene to construct pSEPE.
To make speE overexpression plasmid, PCR was carried out for 25 cycles of 98°C for 0.5 min and 72°C for 3 min with PfuTurbo DNA polymerase (Stratagene) by using primers of sE5ЈNdeI and sE3ЈHind. To make speB overexpression plasmid, PCR was carried out for 25 cycles (94°C for 0.5 min, 55°C for 0.5 min, and 72°C for 1 min) with LA Taq in GC buffer using primers of sB5ЈNdeI and sB3ЈHind (Table II). Each PCR product was digested with NdeI and HindIII and was cloned between the NdeI and HindIII site of expression plasmid pET21cϩ to make pESE8 and pESB8, respectively.
All cloned nucleotide sequences were determined to verify the fidelity of the amplified product to the original sequence using BigDye Terminator Ready Reaction Premix (PerkinElmer Life Sciences) and an ABI PRISM 377 DNA Sequencing System (PerkinElmer Life Sciences).

Gene Disruption
To obtain speB and speE gene homolog disruption strains of T. thermophilus, the thermophile host was genetically transformed as described previously (11,16). All genetic recombinations were confirmed by Southern blot analysis.

HPLC 1
Cells of T. thermophilus were collected from 250 ml of overnight culture (about A 590 ϭ 0.3-0.4) in minimum medium at 70°C. 3-Fold volume of 10% trichloroacetic acid was added to the collected cell paste, and the mixture was vortexed thoroughly to extract cellular polyamines. Appropriately diluted samples were analyzed to determine polyamine composition using a CK-10S column (8.0 ϫ 70 mm, Mitsubishi Kasei) as described previously (17). All samples were subjected to GC and GC-MS analyses to confirm polyamine species, as described by Niitsu et al. (18).

Overexpression and Purification of speE and speB Gene Products
E. coli BL21-CodonPlus(DE3)-RP cells carrying pESE8 or pESB8 were grown in 10 liters of 2ϫ YT medium at 37°C. When apparent absorbance at 600 nm reached 0.3, isopropyl ␤-D-thiogalactopyranoside was added to the culture to a final concentration of 0.5 mM, and incubation was continued for an additional 2 h. For purification of SpeE, cells were collected by centrifugation and suspended in 10 ml of 50 mM Tris-HCl (pH 7.5), 1 mM EDTA followed by cell disruption with a sonic oscillator. After removing the cell debris by centrifugation (30,000 rpm, 20 min), the cell extract was heated for 15 min at 75°C, and the denatured protein was removed by centrifugation (30,000 rpm, 20 min). Heat treatment was repeated twice. The proteins precipitated by 50% saturated ammonium sulfate were collected by centrifugation and suspended in 20 ml of 50 mM Tris-HCl (pH 7.5), 1 mM EDTA. Then the enzyme was further purified by a Hiprep QXL column (Amersham Biosciences) and by Resource HIC PHE column (Amersham Biosciences) chromatographies. SpeE thus obtained was homogeneous on SDS-PAGE and was stored at 4°C as a suspension in 60% saturated ammonium sulfate in 50 mM sodium phosphate buffer (pH 7.0) until use.
For purification of SpeB, cells were corrected and suspended in 10 ml of SB buffer (100 mM HEPES-NaOH (pH 7.5), 5 mM MgCl 2 , 5 mM dithiothreitol, 40 M pyridoxal phosphate, and 1 mM agmatine) and were disrupted by sonication. After removing the cell debris by centrifugation (30,000 rpm, 20 min), the cell extract was heated for 15 min at 75°C, and the denatured protein was removed by centrifugation (30,000 rpm, 20 min). Heat treatment was repeated twice. The supernatant was fractionated by 30 -50% saturated ammonium sulfate. The precipitate collected by centrifugation was suspended in SB buffer. The enzyme was further purified by Hiprep QXL (Amersham Biosciences) column chromatography. Partially purified SpeB was stored at 4°C as a suspension in 60% saturated ammonium sulfate in SB buffer (pH 7.0) until use.

Protein Determination
Protein concentrations were estimated with BCA protein assay reagent kit (Pierce) using bovine serum albumin as a standard.

Enzymatic Reactions
To determine K m and k cat of SpeE, the aminopropyltransferase assay was performed as described in literature (19), except pH was 9.0 and agmatine was used as substrate instead of putrescine. The methyl-14 Clabeled decarboxylated S-adenosylmethionine (dcSAM) was synthesized by the method of Samejima. 2 To confirm the reaction product of SpeE, a reaction mixture consisting of 100 mM Tris-HCl (pH 9.0 at 37°C), 5 mM dithiothreitol, 2 mM agmatine, 337.5 M dcSAM, and 32 g of purified SpeE in a final volume of 2 ml was incubated at 37°C overnight. Polyamines were isolated by using Dowex 50W-X4 (Muromachi Technos Co. Ltd.) column chromatography. Fractions containing polyamines that react with ophthalaldehyde were collected and concentrated.
Alignment of the amino acid sequences of the speB gene products is shown in Fig. 2. The amino acid sequence deduced from the T. thermophilus speB gene homolog showed 29 -32% similarity to those of other organisms. The histidine residue that is critical for catalytic activity in the E. coli enzyme (22) and the manganese ion-binding residues (23) are conserved in all amino acid sequences examined.
To ascertain whether polyamine synthesis starts from ornithine or arginine in the thermophile, we first disrupted the speB gene homolog of T. thermophilus. The disruption strain, named MOSB, was constructed by insertion of the HTK gene into the speB gene homolog of TTY1. When MOSB was cultivated in minimum medium at 70°C, the disruption showed no effect on growth (doubling time of MOSB was 7.2 h while that of TTY1 was 4.9 h; see Fig. 3). Growth on a minimum medium plate at 75°C also showed little difference between MOSB and TTY1 (data not shown). However, disruption of the speB gene homolog resulted in significant defect of growth at 78°C in minimum medium (Table III). In addition, the MOSB colony was pale yellow, whereas the TTY1 colony exhibited a bright yellow color (data not shown).
Intracellular polyamine composition of MOSB grown in minimum medium at 70°C was analyzed by HPLC. Most polyamines found in the cells of TTY1 (Fig. 4A) fell below the limit of detection by our method in MOSB (Fig. 4B). At the same time, an unknown polyamine designated as polyamine A in Fig.  4 was accumulated. Cadaverine, agmatine, and two unknown polyamines (X and Y) were concomitantly detected as minor compounds. Fig. 5A shows the results of GC analysis of polyamines in MOSB. The major peak occurred at 9.3 min, and its molecular weight was estimated as 775 based on mass numbers of molecular fragments observed on the GC-MS spectrum shown in Fig. 5B. This corresponds to the molecular weight of N 1 -aminopropylagmatine bound to three heptafluorobutyryl molecules. Because the gene coding for aminopropyltransferase is present in MOSB, the major unknown polyamine of MOSB can be predicted to be N 1 -aminopropylagmatine. Two additional unknown polyamines (X and Y in Fig. 4B) could not be identified because their amounts were too small.
To confirm the identity of polyamine A, authentic N 1 -amino- propylagmatine was chemically synthesized (Fig. 6). In HPLC analyses, the retention time of the synthesized N 1 -aminopropylagmatine was identical to that of polyamine A (data not shown).
By having identified polyamine A as N 1 -aminopropylagmatine, we next disrupted the gene of polyamine aminopropyltransferase (speE homolog) of MOSB. To construct the speB and speE double-disruption strain, we disrupted the speB homolog by inserting the HTK gene of T. thermophilus (⌬speE) whose speE homolog had been replaced with the pyrE gene. The double disruptant was named T. thermophilus MOSBE. MOSBE showed defective growth at 70°C in minimum medium (doubling time was 12.8 h; see Fig. 3) and significantly defective growth at 78°C (Table III). Like that of MOSB, the MOSBE colony was pale yellow (data not shown). As shown in Fig. 4, agmatine was accumulated in MOSBE cells, and N 1aminopropylagmatine, the major polyamine of MOSB, was drastically diminished. A small amount of cadaverine was detected, but other polyamines, especially long and branched polyamines, were undetectable.
These results indicated that the major polyamine biosynthesis in T. thermophilus starts from arginine and not from ornithine. A new polyamine, N 1 -aminopropylagmatine, produced from agmatine by the speE gene homolog, plays a key role as a metabolic intermediate in polyamine biosynthesis in the thermophile. We therefore concluded that the T. thermophilus speB gene homolog codes for N 1 -aminopropylagmatine ureohydrolase, and the speE gene homolog codes for agmatine aminopropyltransferase.
Although MOSB could not grow in minimum medium at 78°C (Table III), the growth recovered when 250 M spermidine was added to the medium. The polyamine composition of MOSB grown at 78°C is shown in Fig. 7. A long polyamine, homocaldopentamine, was accumulated as the major component in these cells. A small amount of quaternary branched polyamine, tetrakis(3-aminopropyl)ammonium, was also present. N 1 -Aminopropylagmatine was not produced under these conditions, suggesting that the accumulation of long polyamines represses its production. The finding that MOSB produced long and branched polyamines in medium supplemented with spermidine suggests that conversion of N 1 -aminopropylagmatine to spermidine is an essential step in the synthesis of long and branched polyamines and that the speB gene product of T. thermophilus is responsible for the reaction.
To confirm that SpeE converts agmatine to N 1 -aminopropylagmatine and that SpeB uses N 1 -aminopropylagmatine as a substrate to produce spermidine, we purified SpeE and partially purified SpeB to perform in vitro enzymatic reactions. A typical purification procedure of SpeE was summarized in Table IV. SpeE was recovered from the soluble fraction and purified by an anion exchange chromatography and a hydrophobic interaction chromatography. Purity of SpeE by SDS-PAGE was shown in Fig. 8A. The purified SpeE migrated as a single band (36 kDa). The final preparation had a specific activity of 2.4 mol/min/mg proteins, 51-fold that of the crude enzyme.
The purified enzyme was significantly stable; little loss of activity was observed after storage at 4°C in 50 mM sodium phosphate buffer (pH 7.0), 1 M ammonium sulfate for at least a year. The kinetic parameters of SpeE were determined by Hanes-Woolf plot. Because dcSAM decompose at high temperature under alkaline conditions (24), reactions for obtaining kinetic parameters were performed at 37°C (pH 9) with various concentrations of agmatine (0.5-2 M) and 38 M dcSAM. The enzymatic reaction was performed in one tube, and 100 l of reaction mixture was sampled to stop reaction at 0.5, 1, 5, 10, 15, and 20 min. The K m value for agmatine was 0.77 M. The k cat was 0.37 s Ϫ1 when agmatine was used as substrate. The partially purified SpeB migrated as a single band (32.5 kDa) on a SDS-PAGE as shown in Fig. 8B.

DISCUSSION
T. thermophilus produces no fewer than 16 polyamines (Fig.  9C), and its long and branched polyamines have been suggested to play important roles in thermophily (4,8,17). However, the starting point of polyamine biosynthesis of this bacterium remains to be elucidated. We could not find a gene homolog coding for ornithine decarboxylase (speC gene) in the T. thermophilus genome, even though Pantazaki et al. (25) have reported purification and characterization of ornithine decarboxylase of T. thermophilus. Some bacteria such as Rhodopirellula baltica and Selenomonas ruminantium contain a homolog of eukaryotic ornithine decarboxylase with dual specificity on lysine and ornithine (26,27). We could not find a gene homolog coding for eukaryotic ornithine decarboxylase. We speculate two possible explanations for the discrepancy between the report for ornithine decarboxylase and genome sequence data of T. thermophilus: either the gene coding for ornithine decarboxylase has no homology to speC genes from other organisms, or an amino acid decarboxylase, such as arginine decarboxylase or lysine decarboxylase, has broad substrate specificity and accepts ornithine as a substrate.
Phenotypes of Disruption Strains of Polyamine Biosynthetic Genes-In the present study, we constructed an speB gene homolog-disruption mutant of T. thermophilus (MOSB) in order to clarify the first step of polyamine biosynthesis. MOSB exhibited significantly defective growth at 78°C in minimum medium but normal growth at 70°C (Table III and Fig. 3). Polyamines found in MOSB cells (Fig. 4) were cadaverine, agmatine, and N 1 -aminopropylagmatine, compared with 16 or more found in TTY1. In addition, two larger unknown polyamines were found. Judging from their peak positions in HPLC, these compounds may be aminopropylated derivatives of N 1 -aminopropylagmatine. Their identifications will be the subject of future studies. These observations indicate that T. thermophilus can grow at up to 75°C with aminopropylagmatine and other aminopropylated derivatives that could substitute for long and branched polyamines in biochemical reactions at high temperature.
When speB and speE gene homologs of T. thermophilus were disrupted, MOSBE had defective growth even at 70°C (Fig. 3). Agmatine accumulated in MOSBE cells, and the levels of N 1aminopropylagmatine and other polyamines diminished (Fig.  4). A small amount of cadaverine was also detected for the first time in the cells of T. thermophilus. These observations suggest that a gene coding for lysine decarboxylase is present in the T. thermophilus chromosome and is suppressed under normal conditions but is expressed when cellular polyamines disappear, as has been reported for E. coli lysine decarboxylase. In E. coli, expression of the lysine decarboxylase gene is suppressed by putrescine and spermidine (28). Similarly cadaverine may not be present in TTY1 (Fig. 4). It is worthy of note that only agmatine and cadaverine support the growth of the thermophile at lower temperature. Other polyamines, however, are essential for growth at higher temperatures. As shown in Fig.  7, when spermidine was added to the minimum medium, MOSB cells produced long and branched polyamines and were able to survive at 78°C. Therefore, long and branched polyamines are essential for growth at over 75°C. In the wild-type strain, cellular content of these polyamines increased with the rise in the growth temperature (4).
Both MOSB and MOSBE formed pale yellow colonies, although the parent strain, TTY1, forms a bright yellow colony. This suggests that one or more long and/or branched polyamines are required for the following: 1) gene transcription, 2) translation, or 3) enzymatic activity of the enzyme(s) involved in the dye synthesis. The intracellular content of the yellow pigments, carotenoids, which are known act as antioxidants, increase when the cells are grown at higher temperature (29). We speculate that MOSB and MOSBE are unable to grow at over 75°C because of the suppression of carotenoid synthesis and the lack of antioxidant. N 1 -Aminopropylagmatine-In the present study, we identified the major unknown polyamine accumulated in MOSB as N 1 -aminopropylagmatine. A related polyamine in leech, hirudonine (G3-4N), was reported by Robin et al. (30). Chemically synthesized N 1 -aminopropylagmatine also appears in the literature; it was named N 8 -guanylspermidine and is known to be a strong inhibitor of deoxyhypusine synthase (31). However, this is the first time that N 1 -aminopropylagmatine has been found as a natural compound.
Enzymatic Activity of SpeE-When SpeE was incubated with agmatine and dcSAM, only a small amount of aminopropylagmatine was formed as shown in Fig. 8C, trace a. This is not due to low catalytic activity of SpeE protein but is probably due to instability of dcSAM under the conditions employed. The k cat value (0.37 s Ϫ1 ) of the thermophile SpeE is by no means inferior to those of other organisms even at 37°C (19,32). The K m value for agmatine (0.77 M at 37°C) of the present thermophile aminopropyltransferase is smaller than the reported K m value for putrescine (20 M at 37°C) of Thermotoga maritima sper-midine synthase (32). On the other hand, dcSAM is known to be unstable at alkaline pH (24). To reduce the rate of spontaneous degradation of dcSAM, the experiment shown in Fig. 8C, trace a was carried out at 37°C. The reaction mixture was incubated overnight, but the reaction might terminate earlier due to shortage of dcSAM. In the thermophile cells, some mechanisms, such as coupling of reactions, may exist to prevent the degradation of dcSAM at high temperature.
Biosynthetic Pathway of Polyamines in T. thermophilus-Based on the intracellular polyamine composition of MOSB and MOSBE (Fig. 4), we conclude the following: 1) major polyamines are mainly derived from arginine rather than ornithine; 2) SpeE produces N 1 -aminopropylagmatine from agmatine; and 3) SpeB converts N 1 -aminopropylagmatine to spermidine.  a and b), agmatine (traces c and d), or chemically synthesized N 1 -aminopropylagmatine (traces e and f). After 0 and 60 min, reaction mixtures were analyzed by HPLC. Spd, spermidine; Agm, agmatine; aminopropyl-agm, N 1 -aminopropylagmatine. We confirmed that SpeB acts on N 1 -aminopropylagmatine but not on agmatine (Fig. 8C, traces a-f). Sequence homology between "T. thermophilus speB" and speB of other organisms is not high as shown in Fig. 2, and this low similarity would be reflected in the different substrate specificity of the thermophile enzyme. It would be interesting to compare the tertiary structures of SpeB of T. thermophilus and speB gene products of other organisms that differ in substrate specificity. We are currently attempting to crystallize SpeB for structural analyses. The present results suggest the danger of utilizing only sequence homology for genome annotation.
Based on reverse genetic analyses, we propose a new polyamine biosynthetic pathway in T. thermophilus. As shown in Fig. 9B, biosynthesis of major polyamines in T. thermophilus starts only from arginine, which is decarboxylated to form agmatine. An aminopropyl group is added to agmatine by SpeE to form N 1 -aminopropylagmatine. Finally, N 1 -aminopropylagmatine is hydrolyzed to spermidine by SpeB. In this new polyamine biosynthetic pathway, spermidine is synthesized without the production of putrescine. It would be possible to find this polyamine pathway in other organisms.
As shown in Fig. 4, MOSB, which could not produce spermidine from N 1 -aminopropylagmatine, failed to produce long and branched polyamines. Such polyamines were produced in MOSB cells only after addition of spermidine to the minimum medium (Fig. 7). Therefore, the conversion of N 1 -aminopropylagmatine to spermidine is essential for production of long and branched polyamines in T. thermophilus.
Agmatine and aminopropylagmatine were not detected in the wild-type strain (Fig. 4A). The formation of spermidine in wild-type cells may be so rapid that only trace amounts of these intermediates exist. In MOSB cells cultivated in the presence of spermidine, thermospermine should be present as a precursor of homocaldopentamine, but it was not detected by HPLC analysis (Fig. 7). Thermospermine has been identified in the cells of T. thermophilus, as well as homocaldopentamine and other long polyamines. In MOSB, it seems that the aminopropylation of thermospermine is fast enough that only a trace amount of thermospermine is present. These findings suggest that a delicate regulatory network operates for polyamine biosynthesis of T. thermophilus. In the presence of spermidine, production of aminopropylagmatine was suppressed in MOSB (Fig. 7). This finding also suggests the existence of a regulatory system in T. thermophilus.
In the new pathway we have proposed, polyamine synthesis starts with decarboxylation of arginine. Spermidine is produced directly from N 1 -aminopropylagmatine but not from putrescine. The pathway is unique in that putrescine is not involved in the biosynthesis of other polyamines. However, small amounts of putrescine, sym-homospermidine, and other polyamines containing aminobutyl groups are found in wild-type cells, suggesting that genes other than speB are also involved in polyamine metabolism of T. thermophilus. In this context, it is noteworthy that a homolog of a gene coding for deoxyhypusine synthase (33) (HB8, TT0337; HB27, TTC1205) is present in the T. thermophilus genome (both strains HB8 and HB27); and this gene might be involved in sym-homospermidine synthesis. However, not even a trace of sym-homospermidine was detected in MOSB, which accumulated N 1 -aminopropylagmatine in the cells (Figs. 4 and 5). This is consistent with the fact that N 1 -aminopropylagmatine acts as a deoxyhypusine inhibitor (31). To clarify the details of polyamine metabolism in T. thermophilus, further reverse genetic analyses as well as enzymatic investigations are necessary. In addition, detailed properties and structures of SpeB, SpeD, and SpeE are being studied in our laboratory. Structural data for SpeE have been deposited in the Protein Data Base (Protein Data Bank code 1UIR).