1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

N-acetylmuramyl-l-alanine amidases are widely distributed among bacteria. However, in Escherichia coli, only one periplasmic amidase has been described until now, which is suggested to play a role in murein recycling. Here, we report that three amidases, named AmiA, B and C, exist in E. coli and that they are involved in splitting of the murein septum during cell division. Moreover, the amidases were shown to act as powerful autolytic enzymes in the presence of antibiotics. Deletion mutants in amiA, B and C were growing in long chains of unseparated cells and displayed a tolerant response to the normally lytic combination of aztreonam and bulgecin. Isolated murein sacculi of these chain-forming mutants showed rings of thickened murein at the site of blocked septation. In vitro, these murein ring structures were digested more slowly by muramidases than the surrounding murein. In contrast, when treated with the amidase AmiC or the endopeptidase MepA, the rings disappeared, and gaps developed at these sites in the murein sacculi. These results are taken as evidence that highly stressed murein cross-bridges are concentrated at the site of blocked cell division, which, when cleaved, result in cracking of the sacculus at this site. As amidase deletion mutants accumulate trimeric and tetrameric cross-links in their murein, it is suggested that these structures mark the division site before cleavage of the septum.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Cell division in most bacteria calls for the simultaneous constriction of both the cytoplasmic membrane and the murein (peptidoglycan) layer (Rogers et al., 1980). In Gram-negative bacteria, the outer membrane is also invaginated concomitantly (Park, 1996; Nanninga, 1998). Cell septation is normally followed by separation of the daughter cells. This process depends on cleavage of the murein septum that is synthesized during cell division. In most Gram-negative bacteria, the splitting process is already started before completion of septum formation (Burdett and Murray, 1974). This leads to the formation of a V-shaped constriction site characterized by a division furrow, strangling the cell until the two cell halves are detached from one another.

Cleavage of the murein septum has been postulated to be performed by the action of murein hydrolases (Shockman and Höltje, 1994). Indeed, chain-forming mutants isolated from several Gram-positive bacteria that are blocked in the splitting of the septum were found to be deficient in particular murein hydrolases (Tomasz, 1968; Fan, 1970; Forsberg and Rogers, 1974; Shungu et al., 1979; Oshida et al., 1995; Garcia et al., 1999). Surprisingly, in Escherichia coli, the enzymes that cleave the septum have not yet been identified. Quite a number of murein hydrolases with different substrate specificities have been described (Shockman and Höltje, 1994). However, only two classes, namely lytic transglycosylases and endopeptidases (for specificities, see Fig. 1), are known to be capable of cleaving the net-like structure of the murein sacculus (Höltje, 1995). This function is needed to disconnect the two new daughter sacculi during cell division. However, mutants lacking several of the lytic transglycosylases (Lommatzsch et al., 1997) or all the described endopeptidases did not result in the formation of chains of unseparated cells (C. Heidrich, unpublished). Amidases with the specificity of N-acetylmuramyl-l-alanine hydrolases (see Fig. 1) were ruled out as candidates for separating daughter sacculi during cell division in E. coli, as the only amidase described so far was reported to accept exclusively murein degradation products as substrate (van Heijenoort and van Heijenoort, 1971; van Heijenoort et al., 1975; Parquet et al., 1983). This indicates a role in murein recycling but not in division of the sacculus. Later, a gene encoding a 31 kDa amidase was identified at 51 min and named amiA (Tomioka et al., 1983). However, it was never proved that amiA encodes the previously biochemically characterized amidase. Furthermore, the latest map places amiA at 55 min.


Figure 1. Murein chemistry and specificity of the autolytic enzymes of E. coli. GlcNAc, N-acetylglucosamine; MurNAc, N-acetylmuramic acid; A2pm, diaminopimelic acid. Tetrameric muropeptides (not shown) are formed by the addition of a fourth monomer to the free amino group at the A2pm residue of a trimeric cross-bridge.

Download figure to PowerPoint

Additional homologues of amiA have been detected recently in E. coli. Tsui and collaborators detected an open reading frame (ORF) upstream of the mutL repair gene at 94.7 min that encodes a 48 kDa protein with homology to the amidases of Bacillus subtilis, Bacillus licheniformis and a gene product of Salmonella typhimurium (Xu and Elliott, 1993; Quintela et al., 1997). The gene was designated amiB (Tsui et al., 1994). Although it was shown that overexpression of AmiB resulted in lysis, hypersensitivity to osmotic shock conditions and low levels of antibiotics, murein hydrolase activity has not been demonstrated so far. Furthermore, a database comparison for an ORF next to the lytic transglycosylase MltA (Lommatzsch et al., 1997) at 63.4 min revealed a significant similarity (e-values: 10–8) to amidases. The gene encoding a 45 kDa protein was named amiC (GenBank accession no. AAC75856) and, similar to amiB overexpression, resulted in bacteriolysis (M. F. Templin et al., in preparation). Cloning of all three putative amidase genes, amiA, B and C, enabled us to study the specific function of these proteins and to demonstrate that they do indeed encode N-acetylmuramyl-l-alanine amidases. All three amidases turned out to have a major function in septum cleavage during cell division. In addition to their role in cell separation, these amidases can act as powerful autolysins in the presence of murein synthesis inhibitors. Thus, in E. coli, not only lytic transglycosylases and endopeptidases, but also amidases take part in the unrestricted dissolution of the murein sacculus that occurs when antibiotics such as β-lactams block further insertion of new murein subunits into the sacculus (Kitano et al., 1986; Kohlrausch and Höltje, 1991).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Construction of amidase deletion mutants

To obtain more insight into the specific biological function of the amidases in E. coli, we constructed strains lacking one, two or all three amidases. Deletions of amiA and amiB were created by means of polymerase chain reaction (PCR) and introduced into the chromosome using vector pMS7 (see Experimental procedures). The amiC gene was replaced by a kanamycin resistance cassette (M. F. Templin et al., in preparation). To our surprise, except for the amiB mutant MHD 41, the amiA and the amiC mutants were mostly growing in chains of three to six cells (Table 1). In all cases, chaining was most obvious in the late stationary phase. Each possible combination of multiple amidase deletions was then generated by P1 phage transduction from the single knock-out strains. The phenotypes of the double mutants were comparable with that of the single amiC deletion mutant. The double deletion in amiA and amiB showed only a slight tendency to grow in chains (5–10%) when compared with the amiA/C and the amiB/C double mutants, in which about 20% chains were observed. It seems that cell separation could still take place, although at a reduced rate, as long as there was at least one amidase activity left behind. In contrast, almost all the cells in a culture of the mutant MHD 52, lacking all three amidases (Fig. 2A), were growing in long chains of unseparated cells (up to 24 cells).

Table 1.  Morphology of amidase deletion mutants.
MutantDeleted genea(selection marker)Percentage chainsCells/chain
  • a.

    Genes were deleted as described in Experimental procedures.

MHD 8 amiC (kan)20–303–6
MHD 9 amiA (cm)5–103–4
MHD 41 amiB
MHD 44 amiA, amiB (cm)5–103–6
MHD 45 amiA, amiC (cm, kan)30–403–8
MHD 46 amiB, amiC (kan)30–403–8
MHD 52 amiA, amiB, amiC (cm, kan)90–1006–24
MHD 63 amiA, amiB, amiC, sltY (cm, kan)90–1008–40

Figure 2. Scanning electron microscopy of chain-forming amidase mutants of E. coli. The AmiA, B and C triple mutant MHD 52 is shown in (A), and the Slt70, AmiA, B and C mutant MHD 63 in (B). The bar represents 1 µm.

Download figure to PowerPoint

The lytic transglycosylases in E. coli are exoglucosidases, which degrade murein strands in a processive manner commencing at one end of the glycan chain (Romeis et al., 1993). This mode of action would make lytic transglycosylases an ideal candidate to rip up the septum, thereby allowing separation of the daughter sacculi. Hence, we were interested to see whether lytic transglycosylases are also involved in catalysing cell separation. Accordingly, deletion of all three amidases was combined with a deletion of the major lytic transglycosylase Slt70 (Höltje et al., 1975) that, on its own, does not affect cell division (Templin et al., 1992). The mutant, designated MHD 63, formed extremely long chains of up to 40 cells (Fig. 2B). In addition, MHD 63, as well as MHD 52, was also characterized by the presence of ‘minicell’-like dwarf cells within the chains (Fig. 2A). The majority of dwarf cells found in MHD 52 contained DNA as shown by DAPI staining (data not shown).

To prove that the formation of chains of cells was indeed a result of the lack of amidases, complementation of the amidase triple mutant MHD 52 was performed with different pBAD constructs that express the individual amidases AmiA, B and C. In the absence of the inducer arabinose, the mutants containing either of the pBAD plasmids grew with a phenotype comparable with that of the mutants without the plasmid. However, upon addition of 0.001% (w/v) arabinose, the process of separating daughter cells was rescued, and no chains were formed in the cultures any more. Phenotypic complementation of MHD 52 was observed with any of the plasmids carrying amiA, amiB or amiC. In contrast, mutants carrying only the empty vector pBAD continued to grow in long chains regardless of arabinose addition.

Biochemical characterization of AmiA, B and C

The deletion mutant analysis described above indicates an involvement of the three gene products AmiA, B and C in cell separation and, therefore, one would expect these proteins to be active on isolated high-molecular-weight murein sacculi. To characterize the enzymatic specificity of the three putative amidase homologues, the enzymes were overexpressed in the amidase triple mutant MHD 63, which also lacks Slt70, as described in Experimental procedures. In the case of AmiA and AmiC, crude cell fractions were prepared and incubated with murein sacculi. As a control, a crude cell fraction from cells harbouring the empty vector was used. The soluble products released from the sacculi were fractionated by reverse-phase high-pressure liquid chromatography (HPLC) as described in Experimental procedures. The material eluting in a single peak at 10 min in the AmiA and AmiC samples but not in the control was then analysed by mass spectrometry. As shown in Fig. 3, the amidase-specific peak consisted of tri- and tetrapeptides corresponding in mass to the murein peptides Ala-Glu-A2pm and Ala-Glu-A2pm-Ala. Therefore, we conclude that, in both cases, an N-acetylmurmyl-l-alanine amidase activity has been overproduced.


Figure 3. Fractionation and characterization of amidase reaction products. The soluble products released after incubation of murein sacculi with crude cell fractions from AmiA (A) or AmiC (B) overproducing E. coli MHD 63 (pBAD AmiA or pBAD AmiC) (black curve) and control cells harbouring the empty vector (grey curve) were fractionated by reverse-phase HPLC on ODS nucleosil as described in Experimental procedures. The peaks indicated by an arrow were analysed by electrospray ionization mass spectrometry as shown in (A′) and (B′). The fragmentation pattern in the positive ionization mode indicates the presence of the tetrapeptide Ala-Glu-A2pm-Ala (m/z 462.28 Da) and Ala-Glu-A2pm (m/z 391.25 Da).

Download figure to PowerPoint

AmiB, however, formed inclusion bodies that could not be renatured. Therefore, we analysed directly the changes in the murein of the cells that were converted to spheres upon induction of AmiB in the presence of sucrose. With the exception of A2pm-A2pm cross-linked structures, the overall muropeptide pattern (Table 2) was almost unchanged, although the total amount of high-molecular-weight murein sacculi was decreased by about 70% (76% in trial I, 64% in trial II) when calculated on the basis of the UV absorbance. This result is consistent with the action of an amidase that releases peptides and gives rise to the formation of disaccharides after muramidase digestion of the treated murein sacculi. Indeed, new peaks tentatively defined to represent disaccharides, tetra- and tripeptides as well as disaccharide, octa- and disaccharide heptapeptides showed up in the muropeptide elution profile (not shown). Interestingly, the relative amounts of A2pm-A2pm cross-linked muropeptides was increased by about 50% compared with the control (Table 2), indicating that these cross-bridges are poor substrates for AmiB.

Table 2.  Changes in murein structure after overexpression of AmiB in E. coli MHD 63.
MuropeptidesRelative amounts (%)a
ControlAmiBΔ (%)
  1. a. The relative amounts of the muropeptides were calculated as described by Glauner (1988). The values are the means of two independent experiments.

  2. b. Major changes are highlighted in bold figures.

 ld-A2pm-A2pm7.7511.43 +47 .48 b
 ld-A2pm-A2pm4.056.13 +51 .36 b
 ld-A2pm-A2pm0.650.87 +33 .85 b

Although the biochemistry of AmiA, AmiB and AmiC has not been studied in every detail yet, it can be concluded that all three enzymes have N-acetylmuramyl-l-alanine amidase activity and are active with isolated murein sacculi. This finding is in conflict with the biochemical characterization of a mutation that was mapped at 51 min and named amiA (Tomioka et al., 1983), as the mutant showed reduced activity with the monosaccharide derivative MurNAc-L-Ala-D-Glu-m-A2pm, which is no substrate for AmiA deleted in our studies. The discrepancy in the precise mapping position of amiA, 51 min for the mutant studied by Tomioka et al. (1983) and 55 min in the latest E. coli map, respectively, may indicate that it is not the structural gene for AmiA that is affected in that mutant but a regulatory gene that controls the activity of an amidase previously described by van Heijenoort and coworkers (van Heijenoort and van Heijenoort, 1971; van Heijenoort et al., 1975; Parquet et al., 1983). The latter amidase has been shown to prefer MurNAc-L-Ala-D-Glu-m-A2pm as a substrate and to be inactive with isolated sacculi (Parquet et al., 1983). Hence, there seems to be another amidase present in E. coli, the gene of which has not yet been identified.

Murein composition of amidase deletion mutants

The phenotype of the amidase deletion mutants indicated the involvement of amidases in separating the daughter cells during cell division. In the absence of amidase activity, chains of cells with uncleaved septa were formed. Thus, changes in the structure of the murein sacculus are quite likely to occur in the mutants compared with wild-type cells. Therefore, murein sacculi were isolated from the amidase mutants, and the muropeptide composition was determined by reverse-phase HPLC of the muramidase products (Glauner, 1988). None of the amidase single mutants showed a detectable change in any of the muropeptide parameters (Table 3). However, the pronounced morphology of the amidase triple mutant was accompanied by some clear differences in the muropeptide composition in comparison with that of the wild-type strain. The murein of the chain-forming triple mutant showed increased relative amounts of trimers (+53.5%) and tetramers (+155%), as well as a higher percentage of chain ends (+93.5%). (These structures are indicated in Fig. 1.) Conversely, the relative amount of monomers and the average chain length was lower in the mutant. The amount of monomers was decreased by about 9%, whereas the average chain length was 30% shorter. These changes might reflect the defect in the splitting of the septum, as discussed below.

Table 3.  Muropeptide composition of the amidase mutant MHD 52.
MuropeptidesRelative amounts (%)a
MC1061 (WT)MHD 52 (ΔamiA, B, C)Δ (%)
  1. a. and b The relative amounts of the muropeptides, the degree of cross-linkage and the average chain length of the murein strands were calculated as described by Glauner (1988).

Average chain lengthb (disaccharide units)27.519.1−30.5

Electron microscopical analysis of the septa in the chain-forming mutants

Electron microscopy revealed that, in the chain-forming mutants, septa were formed but not cleaved (Fig. 4A). Ingrown murein as well as complete septation of the cytoplasmic membrane was observed, but there was no indication of a cleavage of septal murein. As a consequence of the absence of normal splitting of the murein, invagination of the outer membrane was absent or only marginal. Hence, rather than showing constrictions, the mutants developed septation sites quite reminiscent of Gram-positive bacteria. In addition, incomplete septation sites, where division was just initiated, could be observed frequently (indicated by an arrow in Fig. 4A). These initiation sites gave rise to the formation of ring structures in isolated murein sacculi (Fig. 4B), probably as a result of the increase in murein thickness at these unclipped nascent septation events. Their digestion with murein hydrolases demonstrated that these rings were indeed made of murein (Fig. 4C and D).


Figure 4. Electron microscopy of murein sacculi from E. coli MHD 52.

A. Ultrathin section of a chain of cells. An incomplete septation site is indicated by an arrow.

B. Murein sacculus with a central murein ring.

C. Murein sacculus digested with Cellosyl.

D. Murein sacculus digested with AmiC.

E and F. A hypothetical explanation of the situation given in (C) and (D) respectively. The bar represents 1 µm

Download figure to PowerPoint

Interestingly, there was a significant difference in the fate of the rings when the digestion was performed either by muramidases, such as the lytic transglycosylase Slt70 and Cellosyl, or by peptidases, including the amidase AmiC and the endopeptidase MepA (Keck et al., 1990). The thick murein rings were digested more slowly than the side-wall by Cellosyl as shown in Fig. 4C. After prolonged digestion, however, the rings also disappeared completely. In contrast, the septal murein ring was preferentially attacked by AmiC or MepA, resulting in gaps at these sites (Fig. 5D). This result raises the question whether AmiC and MepA recognize a defined murein structure at the blocked division site. It was tempting to speculate that the observed accumulation of trimeric and tetrameric cross-bridges in the AmiA/B/C deletion mutant (see Table 3) might be the result of an accumulation of these cross-links in the blocked septa (see Fig. 4E and F) and that these are preferentially cleaved by amidases and endopeptidases. However, careful analysis of the changes in the muropeptide composition that occur after digestion of isolated murein sacculi from chain-forming mutants by AmiC did not reveal a specific cleavage of oligomeric cross-bridges or any other particular murein structure (data not shown).


Figure 5. Effect of antibiotics on bacteriolysis of wild type and the amidase mutant. E. coli MC1061 (filled symbols) and MHD 52 (open symbols) were grown in LB medium and incubated with aztreonam (squares) or bulgecin + aztreonam (circles). Arrows indicate the time points of antibiotic addition; fat arrow: bulgecin; thin arrow: aztreonam.

Download figure to PowerPoint

Growth rate and antibiotic resistance of the amidase deletion mutant

The murein hydrolases, by being capable of cleaving covalent bonds in the stress-bearing murein sacculus, represent potentially autolytic enzymes. Indeed, bacteriolysis as a response to cell wall synthesis inhibitors such as penicillins results from the uncontrolled activity of murein hydrolases (Tomasz, 1979). Using the different amidase deletion mutants, we investigated the involvement of the murein amidases in antibiotic-induced cell lysis. As the rate of lysis as a result of inhibition of murein synthesis depends on growth rate (Tuomanen et al., 1986), the growth behaviour of the amidase mutants was determined. Single and double mutants of any of the three amidases were not affected in growth rate, but it was noted that the amidase triple mutant had a prolonged lag phase and reached the end of the logarithmic growth phase at a lower optical density of about OD578 of 2.0 compared with 3.0–3.5 that was reached by the wild type (data not shown). A similar result was also observed for the mutant MHD 63 that lacks Slt70 in addition to AmiA, B and C.

Although penicillin-induced lysis of the amidase mutants was not altered, the involvement of amidases in antibiotic-triggered lysis was revealed by a tolerant response of the amidase triple mutant in the presence of a combination of aztreonam and bulgecin that normally results in bacteriolysis. This peculiar lysis phenomenon was observed some time ago when a mutant with a defect in the major lytic transglycosylase Slt70 was treated with the PBP3-specific β-lactam aztreonam (Templin et al., 1992). The same lytic response can be triggered when wild-type E. coli is treated together with aztreonam and the specific Slt70 inhibitor, bulgecin (Templin et al., 1992). As aztreonam, by specifically blocking septum formation, normally causes E. coli to form stable filaments (Schmidt et al., 1981), autolysis of cells that are deficient in a specific murein hydrolases (Slt70) was a puzzling result. To our excitement, it turned out that lysis does not take place when the amidase triple mutant MHD 52 is treated with a combination of aztreonam and bulgecin (Fig. 5). Wild-type E. coli MC1061 growing in the presence of 4 µg ml−1 bulgecin was lysed when 0.032 µg ml−1 aztreonam was added. In contrast, the mutant MHD 52 did not lyse up to a concentration of 0.25 µg ml−1 aztreonam in the presence of bulgecin. Similarly, the mutant MHD 63, which lacks the three amidases and Slt70, was resistant to aztreonam. This tolerant response of mutants deficient in both amidases and the soluble lytic transglycosylase contrasts with the increased sensitivity of the amidase mutant to aztreonam itself. The minimum inhibitory concentration (MIC) of aztreonam on MHD 52 in the presence of bulgecin was as low as 0.025 µg ml−1, compared with 0.125 µg ml−1 for the wild-type strain. The increase in sensitivity of the amidase mutant towards aztreonam is comparable with the decrease in the MIC that was determined for the Slt70 deletion mutant. Single mutations in the three amidases do not show a tolerant response towards a combined dose of bulgecin and aztreonam. Thus, it seems that the different amidases A, B and C can replace one another. Aztreonam triggers lysis when either all three amidases or the lytic transglycosylase Slt70 are inhibited. However, mutants blocked in all three amidases as well as in the soluble lytic transglycosylase are tolerant against aztreonam. This implies that these hydrolases and the septum-specific transpeptidase PBP3 interact directly with one another. This may be taken as a hint that these enzymes are part of a multienzyme complex, as discussed below.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Like lytic transglycosylases and d,d-endopeptidases, the amidases also come in different flavours (Shockman and Höltje, 1994; Höltje, 1995) and can replace each other in function. With the exception of cytoplasmic AmpD amidase (Höltje et al., 1994; Jacobs et al., 1995), which has a very specific role in recycling murein turnover products and hence a specificity for anhydromuropeptides, it is not clear what kind of function the periplasmic amidases AmiA, B and C have during the cell cycle. A role for amidases in septum separation has been proposed based on the finding that an envA mutation causing chain formation also resulted in reduced amidase activity (Normark et al., 1976;Wolf-Watz and Normark, 1976). However, it was later found that envA encodes the UDP-3-O-acyl-N-acetylglucosamine deacetylase, an enzyme involved in the second enzymatic step of lipid A biosynthesis (Young et al., 1995). Therefore, the formation of chains in envA mutants could also be caused by other changes in the cell wall and not necessarily be directly related to the activity of amidases. The phenotype of the amidase triple mutant MHD52, however, clearly indicates that amidases are important for splitting the septum. Interestingly, the same enzyme specificity is involved in cell separation in a number of Gram-positive bacteria (Tomasz, 1968; Pooley et al., 1972; Forsberg and Rogers, 1974; Chatterjee et al., 1976; Fein and Rogers, 1976; Oshida et al., 1995; Yamada et al., 1996; Garcia et al., 1999).

A rather exciting result is the finding that AmiC or MepA added to isolated murein sacculi of MHD 52 seemed specifically to degrade the thick murein rings resulting from blocked septum formation. It was tempting to speculate that both AmiC and MepA preferentially recognize trimeric cross-bridges that interlink three murein strands, as they are expected to accumulate at sites of uncleaved septa, according to a hypothetical growth model called the ‘three-for-one’ mechanism (Höltje, 1993; 1998). However, a specific cleavage of oligomeric cross-bridges by AmiC could not be demonstrated. Another possibility to explain the observed formation of gaps at the sites of incomplete, uncleaved and, hence, multilayered septa by peptidases but not by muramidases would be a difference in the mechanical stability, predominantly depending on the peptide bridges. It may be that cleavage of peptide bridges in these structures has a more dramatic effect on the overall integrity of the murein netting than it has in the side-wall. One reason could be that higher tensions may occur in the peptide cross-links of a constriction site. An accumulation of oligomeric cross-bridges in uncleaved septa, as speculated above, would be consistent with this interpretation, as it seems feasible to argue that these cross-links are extremely stressed structures.

Although important for cell division, amidases are not essential for cell enlargement. All amidase mutants grow with almost unchanged growth rates and yields under the laboratory conditions tested. Besides playing such an important role during cell division, the amidases are potentially powerful autolytic enzymes and, at least in some cases, are involved in antibiotic-induced bacteriolysis. The puzzling observation that the combined application of aztreonam (a specific inhibitor of PBP3) and bulgecin (a specific inhibitor of the lytic transglycosylase Slt70) causes lysis (Templin et al., 1992) is explained by the tolerant response of the amidase triple mutant. As speculated before, the simultaneous inhibition of PBP3 and Slt70 seems to trigger the activity of a second murein hydrolase. The absence of lysis in the amidase mutant indicates that lysis in the presence of bulgecin and aztreonam is probably caused by the uncontrolled action of amidases. We propose that PBP3, the lytic transglycosylases and the amidases interact directly with one another, probably as part of a multienzyme complex (Höltje, 1996; Vollmer et al., 1999). As depicted in Fig. 6, simultaneous inhibition or deletion of PBP3 and a murein hydrolase, either an amidase or a lytic transglycosylase, may affect the proper architecture of the complex, with the result that, by some allosteric interactions, the second type of murein hydrolase present in the complex is released from its control by the multiprotein complex. In accordance with this speculation, inhibition of the amidases by deletion of the genes and of PBP3 by the addition of aztreonam causes bacteriolysis, probably by triggering the lytic transglycosylase Slt70, as an additional deletion of Slt70 results in a tolerant phenotype. These findings open new strategies to trigger bacteriolysis by specifically interfering with the intimate interaction of the enzymes in the murein-synthesizing machinery.


Figure 6. Schematic representation of the hypothetical effects of a combined deficiency in PBP3 and either the soluble lytic transglycosylase Slt70 or the amidase AmiA, B or C on the functionality of the murein-synthesizing machinery. The murein is represented by open circles (glycan strands) connected by fat lines (peptide cross-bridges). The enzymes are shown by large circles. LT, lytic transglycosylase; AM, amidase; TP, transpeptidase; TG transglycosylase; TP/TG, bifunctional transglycosylase-transpeptidase.

Download figure to PowerPoint

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains and plasmids

The cloning host for the construction of all deletion mutants was E. coli MC1061 (Casabadan and Cohen, 1980). The general cloning vector was pBluescript-II SK+ (Stratagene); the vector pMS7, a sacB-containing modification of pMAK700 (Hamilton et al., 1989), was used for the generation of chromosomal gene knock-outs (Kulakauskas et al., 1991), and pBAD (Invitrogen) was used as expression vector. The kanamycin resistance determinant was taken from pUC4K (Pharmacia); pBCSK+ (Stratagene) was the source of the chloramphenicol resistance marker.

Growth conditions

Bacteria were cultivated aerobically at 37°C in LB medium (Miller, 1972). Agar plates contained 1.5% agar (Gibco BRL). Growth was monitored by optical density readings at 578 nm in an Eppendorf photometer.

Antibiotics were used at the following concentrations: ampicillin (amp), 50 µg ml−1; kanamycin (kan), 50 µg ml−1; chloramphenicol (cm), 20 µg ml−1.

DNA manipulations and PCR

PCR was performed on an MJ Research PTC-200 (Biozym) using 0.5 U 25 µl−1 Powerscript polymerase from Pan Systems to create products with deleted ORFs or using 0.7 U 25 µl−1Taq polymerase (MBI Fermentas) to screen for successful chromosomal gene deletions. Each primer was added to a final concentration of 0.5 µM. After initial denaturation for 3 min at 92°C, touchdown PCR (Don et al., 1991) was performed with 1 min of annealing, 1–3 min of extension, depending on the distance of the primer binding sites, at 72°C and 0.5 min of denaturation at 92°C. The annealing temperature was initially 52°C and then decreased in 12 cycles by 0.5°C in each cycle; finally, the annealing temperature was 46°C for another 12 cycles.

Standard techniques were used to manipulate DNA, and E. coli was transformed using the modified calcium chloride procedure (Sambrook et al., 1989). Restriction endonucleases were purchased from Roche Diagnostics; oligonucleotides came from MWG-Biotech.

Construction of deletion mutants

A modification of the gene exchange method of Kulakauskas et al. (1991) was used to construct a deletion of the coding region of the amidase genes, AmiA and AmiB respectively. First, the respective gene deletions carried by pBluescript-II SK+ were transferred to the vector pMS7. Optionally, a selection marker was introduced at the position of the missing gene using SmaI or BamHI restriction sites created by the primers AmiA/B/C 2 and 3 respectively. For co-integrate formation, growth in liquid culture at 28°C for at least 6 h was followed by plating on LB–amp and incubation at 42°C. Successful co-integrates were found by PCR with the primer combination T7 primer (NEB)–pAmiX5, with X = A or B. Several of the co-integrates were incubated at 28°C in liquid culture for at least 6 h and finally plated on LB plates at 37°C supplemented with 4% sucrose, which is harmful for the pMS7-carrying strains because of the gene locus sacB, and thus selects for curing cells from plasmids and favours replacement of the constructed gene deletion with the chromosomal wild-type copy. The plasmid-borne amiC deletion was transferred to the chromosome of E. coli MC1061 by transduction using Kohara phage 1233 as an intermediate and the technique described by Kulakauskas et al. (1991). Control for successful gene replacement was performed by PCR with the primers pAmiA/B 1 and 5 and for AmiC with pAmiCupup/pUC4Kupup. The transfer of an existing chromosomal deletion into different mutant strains was performed with bacteriophage P1 as described by Miller (1972).

List of primers

AmiA: deletion and control primer


AmiB: deletion and control primer


AmiC: deletion and control primer


Each respective primer pair marked with numbers 1,2 and 3,4 resulted in PCR products upstream and downstream of the selected ORF containing overlapping sequence ends. The probes were mixed giving the final PCR product in a third PCR using primers 1 and 4. The DNA fragment was ligated with pBluescript-II SK+ and used to transform E. coli MC1061. The restriction sites introduced by primers 2 and 3 were used for insertion of the resistance marker.

Overexpression of amidase genes and complementation of amidase deletions

The plasmid pBAD/Myc-HisA (Invitrogen) was used as expression vector. The amidase genes were amplified by PCR with Powerscript polymerase (Pan Systems) using the following primer pairs:

AmiA expression primer


AmiB expression primer


AmiC expression primer


Cloning was carried out according to the manufacturer's recommendations. The resulting vector constructs overexpressing amidase A, B and C were named pBAD AmiA, B and C respectively. For complementation studies, the mutant MHD 63 was transformed with each of the different amidase expression constructs and grown in LB–amp supplemented with 0.8% glucose. Induction was performed in fresh LB–amp with 0.01–0.1% l-arabinose.

Electron microscopy techniques

Bacterial cultures were harvested at an OD578 of about 1.0, resuspended and fixed in 2.5% glutaraldehyde in PBS. The samples were then post-fixed with 1% osmium tetroxide for 1 h on ice and, after rinsing with double-distilled water, treated with 1% aqueous uranyl acetate for 1 h at 4°C. Samples were dehydrated through a graded series of ethanol and embedded in Epon. Ultrathin sections were viewed in a Philips CM10 electron microscope.

For scanning electron microscopy (SEM), the fixed cells were dehydrated in ethanol and critical point dried from CO2. The samples were sputter coated with 8 nm Au/Pd and examined at 20 kV accelerating voltage in a Hitachi S-800 field emission scanning electron microscope.

Purification and enzymatic digestion of murein sacculi for electron microscopy

Murein sacculi of wild-type E. coli and of the amidase mutants were prepared essentially according to the method of de Pedro et al. (1997). Cultures (100 ml) were harvested at an OD578 of about 0.6 by centrifugation at 10 000 g for 5 min. The pellets were resuspended with 3 ml of 0.9 g l−1 NaCl and slowly dropped into 3 ml of boiling 8% (w/v) sodium dodecyl sulphate (SDS). Samples were kept boiling for 4–6 h with stirring and then left at room temperature overnight. After an additional 1 h of boiling, the sacculi were sedimented by ultracentrifugation (Beckman TL100, 400 000 g for 15 min at 30°C) and resuspended with 2.5 ml of 2% (w/v) SDS. After boiling twice for 2 h and sedimentation by ultracentrifugation, the sacculi were resuspended in 2 ml of 50 mM sodium phosphate, pH 7.3. Then, the samples were digested overnight with 200 µg ml−1α-chymotrypsin at 37°C. The incubation was terminated by the addition of 3 ml of 2% (w/v) SDS and boiling of the samples for 4 h in a water bath. Finally, the sacculi were centrifuged again, resuspended in SDS, incubated in a boiling water bath for 2 h and resuspended after another ultracentrifugation step in 100 µl of distilled water. The sacculi were loaded on charged grids according to established methods and used for electron microscopy after fixation with uranyl acetate. Specific digestion of the sacculi with 10 µg ml−1 Cellosyl (kindly provided by Dr Aretz, Aventis, Frankfurt, Germany), Slt70, MepA or Amidase C, prepared as described elsewhere (Keck et al., 1990; Walderich and Höltje, 1991; M. F. Templin et al., in preparation), was performed with the sacculi on the grids.

Analysis of muropeptide composition

Murein sacculi were isolated as described previously (Glauner, 1988). After digestion with α-amylase and pronase, the murein was hydrolysed with Cellosyl. The resulting muropeptides were reduced with sodium borohydride and fractionated by reverse-phase HPLC as described by Glauner (1988).

In vivo effect of overexpressed AmiB on the murein structure

Escherichia coli MHD 63 harbouring pBAD AmiB was grown at 37°C in LB containing 12% sucrose, 10 mM MgSO4, 0.8% glucose and 100 µg ml−1 ampicillin until an OD578 of about 0.4 was reached. An aliquot of the culture was spun down and resuspended in LB medium as above, except that glucose was replaced by 0.1% l-arabinose, and incubated for another 40 min. At this time point, most of the cells in the induced culture were spherical in shape.

The control culture was harvested 15 min before the induced culture. Murein sacculi were prepared and the muropeptide composition was determined according to the established procedure (Glauner, 1988).

Isolation of amidase reaction products

To obtain an enriched fraction of amidase, E. coli MHD 63 harbouring pBAD amiA or pBAD AmiC was grown at 37°C in LB medium. Overexpression of the cloned genes was induced by the addition of 0.1% l-arabinose and continued growth for another 90 min. The soluble cell fraction was prepared at 4°C by breaking the cells in a French press (12 000 psi), followed by centrifugation at 100 000 g for 30 min. In the case of AmiC, the membrane fraction obtained after centrifugation was extracted with 1 M NaCl in 10 mM Tris-maleate, 10 mM MgCl2 buffer, pH 6.8. Murein sacculi (100 µl; 1.3 mg ml−1) were incubated in 10 mM Tris-maleate buffer, pH 5.0 (AmiA) or 9.0 (AmiC), containing 10 mM MgCl2 and 0.1 mM dithioerythritol (DTE) with the crude cell extracts (AmiA, 0.8 mg ml−1; AmiC, 0.2 mg ml−1). The samples were incubated at 32°C (AmiC) or 37°C (AmiA) for 1 h. The reaction was stopped by the addition of 5 µl of 20% phosphoric acid and boiling of the samples for 5 min. Before chromatography, 3 µl of 0.05% trifluoric acid (TFA) was added. The samples (160 µl) were injected onto an ODS Hypersil (5.0 µm) column (125 × 4.6 mm) and eluted at room temperature with a flow rate of 1 ml min−1 using a linear gradient built up in 10 min from 0% buffer A (0.05% TFA in H2O) to 1.6% buffer B (60% acetonitrile in 0.05% TFA–H2O).

Mass spectrometry

The peaks in the HPLC elution profile that were specific for the AmiA and AmiC samples (peaks indicated by an arrow in Fig. 3A and B) were subjected to electrospray ionization mass spectrometry (ESI-MS) (Biemann, 1992) in the positive ionization mode. The dried samples were dissolved in 500 µl of acetonitrile–H2O–acetic acid 50:50:0.05 (v/v/v) and injected at 5 µl min−1 in direct mode into a ESI-Q-TOF mass spectrometer (Micromass) with a capillary voltage of 3687 V, an ion energy of 2.0 V and a desolvation temperature of 140°C. The spectrum in the range from 50 to 1500 Da was monitored for 3 min. The mass spectra (Fig. 3A′ and B′) showed multiprotonated molecules (M + nH)n and a (M + ACN + 2H)2+ peak. To characterize the molecule(s) further, some precursor ions were chosen to be analysed by tandem mass spectrometry (ESI-MS-MS). This allowed us to assign the sequence Ala-Glu-A2pm-Ala to the signal m/z 462.28 Da and Ala-Glu-A2pm to m/z 391.25 Da respectively.

Agar diffusion test and MIC determination

The MIC, defined as the lowest concentration of antibiotic that prevented growth of bacteria, was determined by growth of the bacteria in LB medium at 37°C overnight at 180 r.p.m.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We gratefully acknowledge the encouraging interest in our studies of Uli Schwarz. We express our thanks to Andreas Jaworski (Naturwissenshaftliches und Medizinisches Institut an der Universität Tübingen in Reutlingen) for performing the mass spectrometry. In addition, we thank Yuen-Tsu Nicco Yu for critical reading of the manuscript. The work was supported by a grant from the Deutsche Forschungsgemeinschaft (01KI9704/5).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • 1
    Biemann, K. (1992) Mass spectrometry of peptides and proteins. Annu Rev Biochem 61: 9771010
  • 2
    Burdett, I.D.J. & Murray, R.G.E. (1974) Septum formation in Escherichia coli: characterization of septal structure and the effects of antibiotics on cell division. J Bacteriol 119: 303324
  • 3
    Casabadan, M.J. & Cohen, S.N. (1980) Analysis of gene control signals by DNA fusion and cloning in Escherichia coli. J Mol Biol 138: 179207
  • 4
    Chatterjee, A.N., Wong, W., Young, F.E., Gilpin, R.W. (1976) Isolation and characterization of a mutant of Staphylococcus aureus deficient in autolytic activity. J Bacteriol 125: 961967
  • 5
    Don, R.H., Cox, P.T., Wainwright, B.J., Baker, K., Mattik, J.S. (1991) ‘Touchdown’ PCR to circumvent spurious priming during gene amplification. Nucleic Acids Res 19: 4008
  • 6
    Fan, D.P. (1970) Autolysin(s) of Bacillus subtilis as dechaining enzyme. J Bacteriol 103: 494499
  • 7
    Fein, J.E. & Rogers, H.J. (1976) Autolytic enzyme-deficient mutants of Bacillus subtilis 168. J Bacteriol 127: 14271442
  • 8
    Forsberg, C.W. & Rogers, H.J. (1974) Characterization of Bacillus licheniformis 6346 mutants which have altered lytic enzyme activities. J Bacteriol 118: 358368
  • 9
    Garcia, P., Gonzales, M.P., Garcia, E., Lopez, R., Garcia, J.L. (1999) LytB, a novel pneumococcal murein hydrolase essential for cell separation. Mol Microbiol 31: 12751277
  • 10
    Glauner, B. (1988) Separation and quantification of muropeptides with high-performance liquid chromatography. Anal Biochem 172: 451464
  • 11
    Hamilton, C.M., Aldea, M., Washburn, B.K., Babitzke, P., Kushner, S. (1989) New method for generating deletions and gene replacements in Escherichia coli. J Bacteriol 171: 46174622
  • 12
    Van Heijenoort, Y. & Van Heijenoort, J. (1971) Study of the N-acetylmuramyl-l-alanine amidase activity in Escherichia coli. FEBS Lett 15: 137141
  • 13
    Van Heijenoort, J., Parquet, C., Flouret, B., Van Heijenoort, Y. (1975) Envelope-bound N-acetylmuramyl-l-alanine amidase of Escherichia coli K 12. Purification and properties of the enzyme. Eur J Biochem 58: 611619
  • 14
    Höltje, J.-V. (1993) ‘Three for one’– a simple growth mechanism that guarantees a precise copy of the thin, rod-shaped murein sacculus of Escherichia coli . In Bacterial Growth and Lysis – Metabolism and Structure of the Bacterial Sacculus, Vol. 65. De Pedro, M.A., Höltje, J.-V., and Löffelhardt, W. (eds). New York: Plenum Press, pp. 419426
  • 15
    Höltje, J.-V. (1995) From growth to autolysis: the murein hydrolases in Escherichia coli. Arch Microbiol 164: 243254DOI: 10.1007/s002030050261
  • 16
    Höltje, J.-V. (1996) A hypothetical holoenzyme involved in the replication of the murein sacculus of Escherichia coli. Microbiology 142: 19111918
  • 17
    Höltje, J.-V. (1998) Growth of the stress-bearing and shape-maintaining murein sacculus of Escherichia coli. Microbiol Mol Biol Rev 62: 181203
  • 18
    Höltje, J.-V., Mirelman, D., Sharon, N., Schwarz, U. (1975) Novel type of murein transglycosylase in Escherichia coli. J Bacteriol 124: 10671076
  • 19
    Höltje, J.-V., Kopp, U., Ursinus, A., Wiedemann, B. (1994) The negative regulator of β-lactamase induction AmpD is a N-acetyl-anhydromuramyl-l-amidase. FEMS Microbiol Lett 122: 159164
  • 20
    Jacobs, C., Joris, B., Jamin, M., Klarsov, K., Van Beeumen, J., Mengin-Lecreulx, D. , et al. (1995) AmpD, essential for both beta-lactamase regulation and cell wall recycling, is a novel cytosolic N-acetylmuramyl-l-alanine amidase. Mol Microbiol 15: 553559
  • 21
    Keck, W., Van Leeuwen, A.M., Huber, M., Goodell, E.W. (1990) Cloning and characterization of mepA, the structural gene of the penicillin-insensitive murein hydrolase from Escherichia coli. Mol Microbiol 4: 209219
  • 22
    Kitano, K., Tuomanen, E., Tomasz, A. (1986) Transglycosylase and endopeptidase participate in the degradation of murein during autolysis of Escherichia coli. J Bacteriol 167: 759765
  • 23
    Kohlrausch, U. & Höltje, J.-V. (1991) Analysis of murein and murein precursors during antibiotic-induced lysis of Escherichia coli. J Bacteriol 173: 34253431
  • 24
    Kulakauskas, S., Wikström, P.M., Berg, D.E. (1991) Efficient introduction of cloned mutant alleles into the Escherichia coli chromosome. J Bacteriol 173: 26332638
  • 25
    Lommatzsch, J., Templin, M., Kraft, A., Vollmer, W., Höltje, J.-V. (1997) Outer membrane localization of murein hydrolases: MltA, a third lipoprotein lytic transglycosylase in Escherichia coli. J Bacteriol 179: 54655470
  • 26
    Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press
  • 27
    Nanninga, N. (1998) Morphogenesis of Escherichia coli. Microbiol Mol Biol Rev 62: 110129
  • 28
    Normark, S., Norlander, L., Grundstrom, T., Bloom, G.D., Boquet, P., Frelat, G. (1976) Septum formation-defective mutant of Escherichia coli. J Bacteriol 128: 401412
  • 29
    Oshida, T., Sugai, M., Komatsuzawa, H., Hong, Y.M., Suginaka, H., Tomasz, A. (1995) A Staphylococcus aureus autolysin that has an N-acetylmuramoyl-l-alanine amidase domain and an endo-beta-N-acetylglucosaminidase domain: cloning, sequence analysis, and characterization. Proc Natl Acad Sci USA 92: 285289
  • 30
    Park, J.T. (1996) The murein sacculus. In Escherichia coli and Salmonella, Vol. 1. Neidhardt, F.C., et al. (eds). Washington, DC: American Society for Microbiology Press, pp. 4857
  • 31
    Parquet, C., Flouret, B., Leduc, M., Hirota, Y., Van Heijenoort, J. (1983) N-acetylmuramoyl-l-alanine amidase of Escherichia coli K12. Eur J Biochem 133: 371377
  • 32
    De Pedro, M.A., Quintela, J.C., Höltje, J.-V., Schwarz, H. (1997) Murein segregation in Escherichia coli. J Bacteriol 179: 28232834
  • 33
    Pooley, H.M., Shockman, G.D., Higgins, M.L., Porres-Juan, J. (1972) Some properties of two autolytic-defective mutants of Streptococcus faecalis ATCC 9790. J Bacteriol 109: 423431
  • 34
    Quintela, J.C., De Pedro, M.A., Zollner, P., Allmaier, G., Garcia-del Portillo, F. (1997) Peptidoglycan structure of Salmonella typhimurium growing within cultured mammalian cells. Mol Microbiol 23: 693704
  • 35
    Rogers, H.J., Perkins, H.R., Ward, J.B. (1980) Microbial Cell Walls and Membranes. London: Chapman & Hall
  • 36
    Romeis, T., Vollmer, W., Höltje, J.-V. (1993) Characterization of three different lytic transglycosylases in Escherichia coli. FEMS Microbiol Lett 111: 141146
  • 37
    Sambrook, J., Fritsch, E.F., Maniatis, T. (1989) Molecular Cloning. A Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press
  • 38
    Schmidt, L.S., Botta, G., Park, J.T. (1981) Effects of furazlocillin, a beta-lactam antibiotic which binds selectively to penicillin-binding protein 3, on Escherichia coli mutants deficient in other penicillin-binding proteins. J Bacteriol 145: 632637
  • 39
    Shockman, G.D. & Höltje, J.-V. (1994) Microbial peptidoglycan (murein) hydrolases. In Bacterial Cell Wall. Ghuysen, J.-M., and Hakenbeck, R. (eds). Amsterdam: Elsevier Science, pp. 131166
  • 40
    Shungu, D.L., Cornet, J.B., Shockman, G.D. (1979) Morphological and physiological study of autolytic-defective Streptococcus faecium strains. J Bacteriol 138: 598608
  • 41
    Templin, M.F., Edwards, D.H., Höltje, J.-V. (1992) A murein hydrolase is the specific target of bulgecin in Escherichia coli. J Biol Chem 267: 2003920043
  • 42
    Tomasz, A. (1968) Biological consequences of the replacement of choline by ethanolamine in the cell wall of pneumococcus: chain formation, loss of transformability, and loss of autolysis. Proc Natl Acad Sci USA 59: 8693
  • 43
    Tomasz, A. (1979) The mechanism of the irreversible antimicrobial effects of penicillins: how the betalactam antibiotic kill and lyse bacteria. Annu Rev Microbiol 33: 113137
  • 44
    Tomioka, S., Nikaido, T., Miyakawa, T., Matsuhashi, M. (1983) Mutation of the N-acetylmuramyl-l-alanine amidase gene of Escherichia coli. J Bacteriol 156: 463465
  • 45
    Tsui, H.-C.T., Zhao, G., Feng, G., Leung, H.-C.E., Winkler, M.E. (1994) The mutL repair gene of Escherichia coli K-12 forms a superoperon with a gene encoding a new cell-wall amidase. Mol Microbiol 11: 189202
  • 46
    Tuomanen, E., Cozens, R., Tosch, W., Zak, O., Tomasz, A. (1986) The rate of killing of Escherichia coli by beta-lactam antibiotics is strictly proportional to the rate of bacterial growth. J Gen Microbiol 132: 12971304
  • 47
    Vollmer, W., Von Rechenberg, M., Höltje, J.-V. (1999) Demonstration of molecular interactions between the murein polymerase PBP1B, the lytic transglycosylase MltA, and the scaffolding protein MipA of Escherichia coli. J Biol Chem 274: 67266734DOI: 10.1074/jbc.274.10.6726
  • 48
    Walderich, B. & Höltje, J.-V. (1991) Subcellular distribution of the soluble lytic transglycosyalse in Escherichia coli. J Bacteriol 173: 56685676
  • 49
    Wolf-Watz, H. & Normark, S. (1976) Evidence for a role of N-acetylmuramyl-l-alanine amidase in septum separation in Escherichia coli. J Bacteriol 128: 580586
  • 50
    Xu, K. & Elliott, T. (1993) An oxygen-dependent coproporphyrinogen oxidase encoded by the hemF gene of Salmonella typhimurium. J Bacteriol 175: 49904999
  • 51
    Yamada, S., Sugai, M., Komatsuzawa, H., Nakashima, S., Oshida, T., Matsumoto, A., Suginaka, H. (1996) An autolysin ring associated with cell separation of Staphylococcus aureus. J Bacteriol 178: 15651571
  • 52
    Young, K., Silver, L.L., Bramhill, D., Cameron, P., Eveland, S.S., Raetz, C.R. , et al. (1995) The envA permeability/cell division gene of Escherichia coli encodes the second enzyme of enzyme of lipid A biosynthesis. J Biol Chem 270: 3038430391DOI: 10.1074/jbc.270.51.30384