Cell wall structural divergence among Thermus spp.


  • José Carlos Quintela,

    1. Centro de Biologı́a Molecular “Severo Ochoa”, Consejo Superior de Investigaciones Cientı́ficas-Universidad Autónoma de Madrid, Facultad de Ciencias, Campus de Cantoblanco, 28049 Madrid, Spain
    Search for more papers by this author
  • Peter Zöllner,

    1. Institute for Analytical Chemistry, University of Vienna, Währinger Strasse 38, A-1090 Vienna, Austria
    Search for more papers by this author
  • Francisco Garcı́a-del Portillo,

    1. Centro de Biologı́a Molecular “Severo Ochoa”, Consejo Superior de Investigaciones Cientı́ficas-Universidad Autónoma de Madrid, Facultad de Ciencias, Campus de Cantoblanco, 28049 Madrid, Spain
    Search for more papers by this author
  • Günter Allmaier,

    1. Institute for Analytical Chemistry, University of Vienna, Währinger Strasse 38, A-1090 Vienna, Austria
    Search for more papers by this author
  • Miguel A de Pedro

    Corresponding author
    1. Centro de Biologı́a Molecular “Severo Ochoa”, Consejo Superior de Investigaciones Cientı́ficas-Universidad Autónoma de Madrid, Facultad de Ciencias, Campus de Cantoblanco, 28049 Madrid, Spain
    Search for more papers by this author

*Corresponding author. Tel.: +34 (1) 3978083; Fax: +34 (1) 3978087; E-mail: madepedro@cbm.uam.es


The fine structure of sacculi from Thermus thermophilus HB27, T. aquaticus YT-1 and Thermus ATCC27737 has been worked out by HPLC analysis and mass spectrometry techniques. The three microorganisms have a murein composition of the rare A3β chemotype, but showed substantial differences in muropeptide composition. Phenylacetylated muropeptides, previously described in T. thermophilus HB8, were detected exclusively in T. thermophilus HB27. Murein from T. aquaticus YT-1 was devoid of d-Ala-d-Ala terminated muropeptides, which were, in contrast, abundant in T. thermophilus HB27 and Thermus ATCC27737. The significance of these findings is discussed.


The peptidoglycan (murein) sacculus constitutes the primary morphogenetic element of the bacterial cell [1]. Although murein basic architecture is highly conserved among bacteria, a large number of variations have been identified [2, 3]. Variations in murein structure likely reflect evolutionary adaptations to particular environments, and have important taxonomic implications [3].

Thermus spp. are extreme thermophilic bacteria living in hot springs [4]. Together with the genus Deinococcus form one of the earlier branches in bacterial phylogenetic trees [5]. Investigation of Thermus thermophilus HB8 cell wall revealed a unique murein structure [6]. First, the primary structure (monomeric subunit: N-acetylglucosamine-N-acetylmuramyl-l-Ala-d-Glu-(γ)-l-Ornithinyl [(δ)Gly-Gly]-d-Ala-d-Ala) corresponded to the rare A3β chemotype only found before in the phylogenetically related species Deinococcus radiodurans[3, 7]. Second, about 25% of total muropeptides have the (δ)NH2 of ornithine (Orn) acylated by phenylacetyl-Gly instead of (Gly)2. T. thermophilus HB8 is the only known case with phenylacetate (PA) as a murein component. The extreme rarity of T. thermophilus HB8 murein structure, led us to investigate whether or not its distinctive features were conserved among other members of the genus, as they could represent important taxonomical traits for the genus Thermus. Therefore murein from representative strains of the species T. thermophilus and T. aquaticus, and from the isolate Thermus ATCC27737 was investigated.

2Materials and methods

2.1Bacteria and growth conditions

T. thermophilus HB27 [8], T. aquaticus YT-1 (ATCC25104) and Thermus sp. ATCC27737 were used in this work. Bacteria were grown at 70°C in 4 g l−1 yeast extract, 8 g l−1 Bacto Peptone, and 3 g l−1 of NaCl, pH 7.5 [9], except T. aquaticus YT-1 which was grown in the same medium diluted to one half.

2.2Murein preparation and HPLC analyses

Bacterial suspensions were dropped onto an equal volume of boiling 8% sodium dodecylsulfate and kept for 1 h under magnetic stirring. Detergent was removed by repeated cycles of centrifugation and resuspension in water. Murein samples were muramidase (Cellosyl, Hoechst, Germany) digested and analysed by HPLC on a Hypersil ODS 18 reverse-phase column (250×4 mm, 3 mm particle size) (Teknochroma, Barcelona, Spain) [6, 10, 11]. Muropeptides were detected by monitoring A204. For amino acid analysis murein samples were hydrolysed in 6 N HCl for 12 h at 105°C, vacuum dried, resuspended in water, and further processed for ortho-phtaldialdehide pre-column derivatisation and HPLC analysis as described [6].

2.3HPLC analysis of muropeptide-derived lactylpeptides

Samples (500 μg) of unreduced muropeptide mixtures were subjected to β-elimination with 4 M [NH4]OH for 1 h at 37°C [12]. The mixtures were vacuum dried, resuspended in water, and analysed by HPLC on a Hypersil RP 18 column (3 μm of particle size, 250×4 mm) at room temperature. Column was eluted at 0.5 ml min−1 for 5 min with 0.05% trifluoracetic acid (TFA) in ultrapure water (MilliQ system, Millipore), and then with a 115 min linear gradient to 30% acetonitrile, 0.05% TFA in ultrapure water. Eluent was monitored on an UV detector set at 204 or 258 nm. Desalting of lactylpeptide mixtures was performed by HPLC as above except for the elution program: 8 min 0.05% TFA in water, 20 min linear gradient to 30% acetonitrile, 0.05% TFA and 15 min isocratic in 30% acetonitrile, 0.05% TFA. Lactylpeptides eluted between 18 and 32 min and were collected as one fraction (7 ml) which was lyophilised and subjected to further analysis.

2.4Matrix-assisted laser desorption/ionisation time-of-flight (MALDI-TOF) mass spectrometry

Measurements were performed on a Kompact MALDI III (Shimadzu Kratos Analytical, Manchester, UK) laser desorption time-of-flight instrument equipped with a type VSL-337ND (Laser Science, Newton, MA, USA) nitrogen laser (λ=337 nm, 3 ns pulse width). Positive ions were accelerated from the target in the continuous mode to a final potential of 20 kV and laser power density was kept near threshold level. All spectra were acquired in the reflector mode. Mass spectra were obtained by signal averaging of 100 laser shots and then smoothed by Savitsky-Golay algorithm. Calibration of the instrument was done externally with the [M+H]+ ions of the applied matrix and the [M+H]+ ions of standard peptides (Leu-Trp-Met-Arg-Phe-Ala, angiotensin III, substance P and bee venom mellitin). Average mass values are reported. Sample preparation was as reported in [13] with α-cyano-4-hydroxy-trans-cinnamic acid matrix (Sigma, St. Louis, MO, USA). 0.5 μl of the matrix solution (20 mg ml−1 acetone) was placed in the centre of each sample well of the disposable stainless-steel 20-well target and the solvent was evaporated quickly at room temperature. The desalted mixture of lactylpeptides was redissolved in 50 μl of water containing 0.1% TFA. Sample solution (0.5 μl) was placed on top of the matrix surface, dried and inserted into the ion source.


3.1Muropeptide composition of murein from Thermus spp.

Amino acid analysis of muramidase (Cellosyl Hoechst, Frankfurt am Main, Germany) digested sacculi from T. thermophilus HB27, Thermus ATCC27737, and T. aquaticus YT-1 revealed Glu, Orn, Ala and Gly as the only amino acids in the muramidase-solubilised material. In all instances less than 3% of the total amount of Orn, remained insoluble after muramidase treatment, indicating a satisfactory solubilisation of the mureins.

Murein samples (200 μg) were muramidase digested, and the solubilised fractions were analysed by HPLC. The corresponding chromatograms are shown in Fig. 1. The elution profile for T. thermophilus HB27 was virtually identical to the one established previously for the HB8 strain, including PA containing muropeptides [6]. Coincidence in the retention times (Rt's) of peaks was confirmed running a sample of T. thermophilus HB8 murein as reference (data not shown). Therefore peaks with the same Rt's in T. thermophilus HB27 and HB8 were considered identical (Fig. 1). Elution profiles for T. aquaticus YT-1, and Thermus ATCC27737 were quite different to T. thermophilus. The more distinctive features were the absence of PA-muropeptides in T. aquaticus YT-1 and Thermus ATCC27737, and the reduction in the number of major peaks in T. aquaticus YT-1 (Fig. 1). Absence of PA-muropeptides was confirmed in HPLC runs with the detector set at 258 nm, where PA but not normal muropeptides absorb strongly. All the significant peaks of T. aquaticus YT-1, and Thermus ATCC27737 eluted with Rt's coincident with muropeptides of T. thermophilus. Lactylpeptide HPLC elution profiles were specific for each strain, although peaks in the simpler chromatograms had equivalents with identical Rt's into the more complex ones (Fig. 2). Such results suggested that mureins from T. aquaticus YT-1, and Thermus ATCC27737 were made up of subsets of T. thermophilus muropeptides. PA-acylated lactylpeptides, identified for their absorption at 258 nm, were found only in T. thermophilus HB27 (Fig. 2).

Figure 1.

HPLC analysis of murein samples. Murein from (A) T. thermophilus HB27, (B) Thermus ATCC27737 and (C) T. aquaticus YT-1 were subjected to HPLC analysis. Numbers in the chromatogram for T. thermophilus HB27 correspond to: 1, l-Orn[(α)-R]-d-Ala; 2, (Gly)2-(δ)-l-Orn[(α)-R]-Gly; 3, l-Orn[(α)-R]-(d-Ala)2+Gly-(δ)-l-Orn[(α)-R]-d-Ala; 4, Gly-(δ)-l-Orn[(α)-R]-d-Ala-Gly; 5, (Gly)2-(δ)-l-Orn[(α)-R]-d-Ala; 6, (Gly)2-(δ)-l-Orn[(α)-R]-d-Ala-Gly; 7, Gly-(δ)-l-Orn[(α)-R]-(d-Ala)2; 8, (Gly)2-(δ)-l-Orn[(α)-R]-(d-Ala)2; 9, Gly-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-d-Ala; 10, (Gly)2-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-d-Ala+(Gly)2-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-d-Ala-Gly; 11, Gly-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-(d-Ala)2; 12, (Gly)2-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-(d-Ala)2; 13, (Gly)2-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-d-Ala; 14, (Gly)2-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-R]-(d-Ala)2; 15, PA-Gly-(δ)-l-Orn[(α)-R]-Gly+Gly-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-(Anh)R]-(d-Ala)2; 16, (Gly)2-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-(Anh)R]-(d-Ala)2; 17, PA-Gly-(δ)-l-Orn[(α)-R]-d-Ala+PA-Gly-(δ)-l-Orn[(α)-R]-d-Ala-Gly; 18, PA-Gly-(δ)-l-Orn[(α)-R]-(d-Ala)2; 19, PA-Gly-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-(Anh)R]-d-Ala+PA-Gly-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-(Anh)R]-d-Ala-Gly; 20, PA-Gly-(δ)-l-Orn[(α)-R]-d-Ala→(Gly)2-(δ)-l-Orn[(α)-(Anh)R]-(d-Ala)2. Abbreviations: R, N-acetylglucosaminyl-(β 1→4)-N-acetylmuramyl-l-Ala-d-Glu-(γ)→; (Anh)R, N-acetylglucosaminyl-(β 1→4)-[(1→6) anhydro]-N-acetylmuramyl-l-Ala-d-Glu-(γ)→; PA, phenylacetate. Numbers in the chromatograms for Thermus ATCC27737 and T. aquaticus YT-1, identify the corresponding muropeptides with respect to T. thermophilus HB27, assuming identity of peaks with equal retention times.

Figure 2.

HPLC analysis of muropeptide-derived lactylpeptides. Muropeptide mixtures from (A) T. thermophilus HB27, (B) Thermus ATCC27737 and (C) T. aquaticus YT-1 were subjected to β-elimination and analysed by HPLC. Eluate was monitored either at 204 or at 258 nm in all instances. Samples B and C did not contain muropeptides detectable at 258 nm. Arrowheads indicate the lactylpeptides common to all samples. Range for Y-axis was the same at 204 and 258 nm.

3.2MALDI-TOF mass spectrometry of lactylpeptide mixtures

Mixtures of lactylpeptides from T. aquaticus YT-1 and Thermus ATCC27737 were analysed by MALDI-TOF to identify individual molecular species present in the mixtures. Sodium-carrying molecular ions of the major lactylpeptides were detected as dominating ions, although more complex sodium and/or potassium adduct ions —?— [M+K]+, [M+Na+K−H]+, [M+2Na+K−2H]+—?— were also formed (Fig. 3). The m/z values measured for the different quasimolecular ions were compared in Table 1 with values calculated for lactylpeptides derived from muropeptides of T. thermophilus[6]. Three identical lactylpeptides were identified in both species (Table 1) which corresponded to the peptide moieties of muropeptides 4 plus 5, 10a and 13 of T. thermophilus (Fig. 1) [6]. These results supported a common amino acid sequence for the peptide side chains of mureins from T. aquaticus YT-1, Thermus ATCC27737 and T. thermophilus, and indicated that muropeptides with equal Rt's in the HPLC elution patterns were indeed the same for the three microorganisms.

Figure 3.

MALDI-MS analysis of lactylpeptides from (A) T. aquaticus YT-1, and (B) Thermus ATCC27737. Aliquots of the lactylpeptide mixtures used for HPLC analysis were subjected to MALDI-MS. The series of peaks for each single lactylpeptide are grouped by lines. The numbers indicate the peaks in Fig. 1 able to release the corresponding lactylpeptide.

Table 1.  Calculated and measured m/z values for the positive quasimolecular ions of major lactylpeptides
Lactylpeptidec (Peak in Fig. 1A)dIon typeT. aquaticus YT-1Thermus ATCC27737
  Calculated m/zMeasured m/zΔmaError (%)bMeasured m/zΔmaError (%)b
  1. aMass difference between measured and calculated quasimolecular ion values.

  2. b[(measured mass−calculated mass)/calculated mass]×100.

  3. cR, d-lactyl-l-Ala-d-Glu-(γ)→.

  4. dSource muropeptides in Fig. 1A for the indicated lactylpeptides.

Gly-(δ)-l-Orn[(α)-R]-d-Ala-Gly+Gly-Gly-(δ)-l-Orn[(α)-R]-d-Ala (4+5)[M+Na]+612.60612.74+0.14+0.02612.74+0.14+0.02
Gly-Gly-(δ)-l-Orn[(α)-R]-d-Ala→Gly-Gly-(δ)-l-Orn[(α)-R]-d-Ala (10)[M+Na]+1184.191184.73+0.54+0.051184.73+0.54+0.05
Gly-Gly-(δ)-l-Orn[(α)-R]-d-Ala→Gly-Gly-(δ)-l-Orn[(α)-R]-d-Ala→Gly-Gly-(δ)-l-Orn[(α)-R]-d-Ala (13)[M+Na]+1755.761756.65+0.89+0.051756.65+0.89+0.05

3.3Muropeptide proportions of Thermus spp. mureins

Muropeptide proportions were calculated on the assumption that HPLC peaks with equal Rt's correspond to identical muropeptides in all samples. An automatic peak-integrator (Spectra Physics 4270, Spectra-Physics, San Jose, CA, USA) was used and integration areas were corrected for the extinction coefficient of individual muropeptides as stated by Glauner [11], and Quintela [6]. The results are summarised in Table 2. More remarkable features were the apparent lack of muropeptides with d-Ala-d-Ala C-terminal dipeptides in T. aquaticus YT-1 (peaks 7, 8, 11, 12 in Fig. 1) and the presence of PA-muropeptides in T. thermophilus HB27 only. In all instances dimers were the major cross-linked muropeptides with sizable proportions of trimers but no higher order oligomers were detected.

Table 2.  Muropeptide composition of murein from Thermus spp.
MuropeptideaRelative abundance (mol %)
 T. thermophilus HB27Thermus ATCC27737T. aquaticus YT-1
  1. aNumbers correspond to peaks in Fig. 1.

  2. bTwo muropeptides coelute (see legend to Fig. 1).

  3. cMuropeptides with d-Ala-d-Ala at the C-terminal of the peptide side chain. As no other d-Ala-d-Ala terminated muropeptide was found in T. aquaticus YT-1, it was assumed that the d-Ala-d-Ala muropeptide eluting in peak 3 was also absent.

  4. dMuropeptides with d-Ala-Gly at the C-terminal of the peptide side chain.

  5. eMuropeptides with phenylacetyl-Gly bound to the ornithine δ-NH2 group.

  6. n.d., not detected.



Muropeptide analysis of murein from T. aquaticus YT-1, Thermus ATCC27737, and T. thermophilus HB27 (found to be virtually identical to strain HB8 [6]) demonstrated a substantial diversity in muropeptide composition. Nevertheless, HPLC results were congruent with a common murein basic architecture for the three Thermus spp. This was confirmed by the positive identification of identical muropeptides by MALDI-TOF mass spectrometry of lactylpeptide mixtures. Therefore it was concluded that the three mureins conformed to the rare A3β chemotype, as T. thermophilus HB8, with N-acetylglucosaminyl-N-acetylmuramyl-l-ala-d-Glu-(γ)-l-Orn[(δ)-Gly-Gly]-d-Ala-d-Ala as the basic monomer, and (Gly)2 mediated cross-linking [3, 6].

Muropeptide compositions were largely different because of variability in the proportion of d-Ala-d-Ala terminated muropeptides, and because PA-muropeptides were only present in T. thermophilus. d-Ala-d-Ala muropeptides accounted for about 50% of the total in T. thermophilus, but could not be detected in T. aquaticus YT-1. The disparity observed could reflect a deficiency in DD-carboxypeptidase activity in T. thermophilus as demonstrated for Caulobacter crescentus[14]. In contrast, d-Ala-Gly muropeptides were abundant in all the mureins tested (25 to 34 mol %). d-Ala-Gly muropeptides have been found as trace components in Escherichia coli[8], and in higher proportions in C. crescentus and D. radiodurans[7, 14]. In E. colid-Ala-Gly muropeptides are generated by d-Ala-d-Ala ligase miss-incorporation of Gly [10, 15]. Whether the same mechanism applies to bacteria rich in such muropeptides is not known. Interestingly T. aquaticus YT-1 murein was devoid of d-Ala-d-Ala muropeptides but was rich in d-Ala-Gly muropeptides, indicating that the corresponding DD-carboxypeptidase(s) might be highly specific for d-Ala-d-Ala muropeptides.

According to our results accumulation of PA-muropeptides is a distinct characteristic of T. thermophilus. The fact that T. thermophilus murein had the higher proportion of d-Ala-d-Ala muropeptides but the lower cross-linkage, suggests that PA-muropeptides, blocked as acceptors for transpeptidation, could modulate cross-linkage in T. thermophilus. However the previously proposed implication of PA-muropeptides in thermostability is seriously question by their absence from mureins of T. aquaticus and ATCC27737, which have growth temperature ranges similar to T. thermophilus[6].

The structural differences observed in Thermus spp. cell walls imply deep adaptations in murein metabolism. From an evolutive point of view the unique ability of T. thermophilus to accumulate PA-muropeptides is most interesting. Synthesis of PA-muropeptides is likely to require a set of dedicated enzymes which could have been in the common ancestor and conserved only in T. thermophilus, or alternatively could have been gained through T. thermophilus speciation. In any case, the divergence in the ability to synthesise PA-muropeptides is likely to correspond to some decisive event in the evolutionary history of the genus Thermus.


The technical assistance of J. de la Rosa is greatly appreciated. This work was supported by grant BIO94-0789 from the CICYT and an institutional grant from ‘Fundación Ramón Areces’ to M.A.P., and by grant no. 11183 from the Austrian Science Foundation to G.A. Travel money was provided by ‘Acción Integrada Austria-España 22B’. J.C.Q. was supported by a fellowship from Fundación Rich, Madrid.