Module 4.7.1.5

Peptidoglycan Recycling

TSUYOSHI UEHARA^† AND JAMES T. PARK*

[SECTION EDITOR, LYNN SILVER]

Posted October 15, 2008

Department of Molecular Biology and Microbiology, Tufts University School of Medicine, Boston, MA 02111

*Corresponding author. Mailing address: Department of Molecular Biology and Microbiology, Tufts University School of Medicine, Boston, MA 02111. Phone: (617) 636-6753, Fax: (617) 636-0337, E-mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it

^†Present address: Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, MA 02115

HISTORICAL PERSPECTIVE

Peptidoglycan (PG) recycling is a process whereby most of the PG of the side walls of the cell is degraded and all of its components are recovered and made available to the cell to synthesize more PG or to use as an energy source. This phenomenon was totally hidden from view until relatively recently. It is now of interest because of its relationship to β-lactamase induction. In 1985 it was known that Bacillus subtilis (31), Lactobacillus acidophilus (2), and Neisseria gonorrhoeae (16, 19), all of which contain meso-diaminopimelic acid (Dap) in their PG, lost 25 to 50% of the Dap from their cell wall to the medium each generation, whereas Escherichia coli lost less than 6% (4, 31). Goodell apparently wondered about the contrast between E. coli and the other organisms and reexamined turnover by E. coli. He confirmed that only a relatively small amount (6 to 8%) was being lost and showed that the lost Dap could be recovered in the spent medium as the peptides l-Ala-γ-d-Glu-Dap, l-Ala-γ-d-Glu-Dap-d-Ala, and Dap-d-Ala (18). He then performed a pulse-chase experiment which unexpectedly revealed that much of the Dap from PG was being recycled each generation instead of being lost to the medium (18). For the experiment, E. coli lys dap was labeled with ³H-Dap for two generations and then chased, and the cell's contents were analyzed (18). The key finding was that the radioactivity of UDP-N-acetylmuramic acid (MurNAc)-l-Ala-γ-d-Glu-Dap-d-Ala-d-Ala (UDP-MurNAc-pentapeptide), a precursor of PG, remained high, while that of the Dap pool dropped almost 100-fold. The ³H-Dap required to maintain the high radioactivity in the UDP-MurNAc-pentapeptide pool could only have come from the PG, and therefore recycling must be occurring. From the data it could be calculated that 45% of the Dap was being recycled each generation (15). It was subsequently shown that l-Ala-γ-d-Glu–³H-Dap could be incorporated directly into PG and that OppA, the periplasmic binding protein for oligopeptide permease, was required for uptake (15, 17). This latter conclusion proved incorrect in that a specific periplasmic binding protein for the murein tripeptide, MppA, was required for uptake of l-Ala-γ-d-Glu-Dap (43). However, 9 years later it was discovered that even in the absence of oligopeptide permease, recycling proceeded normally (41). This was followed by the discovery that AmpG (required for β-lactamase induction) is a permease essential for PG recycling (24). In the past dozen years since the discovery of the role of AmpG, 11 more enzymes specific for recycling have been identified and are discussed below.

OVERVIEW OF PG RECYCLING

As mentioned earlier, it has been estimated that 45% of the cell wall Dap is recycled each generation (15). The turnover is confined to the side walls, since the poles of the cell are stable (3, 11, 12). This indicates that over 60% of the side wall is actually being degraded and recycled each generation. The reason for this turnover is completely unknown but seems not to be related to the cell wall elongation process. For cell elongation, new strands of PG are inserted one strand at a time between existing strands to elongate the sacculus (8, 10, 39).

Eleven genes (of those listed in Table 1 with the exception of nagA and amiD) have been identified in E. coli which appear to have as their sole function the recovery of degradation products from PG, thereby making them available to the cell for synthesizing more PG or to use as an energy source. Since PG represents only 2.5% of the cell mass of E. coli (37), it does not represent a large energy source. However, faced with a sudden loss of carbon or energy, recycling may provide the necessary building blocks to complete a round of cell division and enter stationary phase.

Table 1Murein recycling enzymes

What is known is that recycling is not essential under laboratory conditions. E. coli cells lacking AmpG permease (essential for recycling) grow well in minimal media. With the exception of ldcA, mutations in the other 12 known genes involved in recycling (Table 1) have no phenotype. All of the enzymes are produced and are active during log growth. All are active on substrates that are present at concentrations in the 10⁻⁵ molar range.

As shown in Table 1, with the exception of the AmpG permease located in the inner membrane, MppA located in the periplasm, and AmiD located in the outer membrane, all the enzymes are found in the cytoplasm. It is interesting that AmpD, LdcA, and MpaA cleave bonds in 1,6-anhydro-N-acetylmuramic acid (anhMurNAc)-l-Ala-γ-d-Glu-Dap-d-Ala, a recycling intermediate, without cleaving similar bonds in the essential PG precursor, UDP-MurNAc-pentapeptide. This allows the processes of PG biosynthesis and PG recycling to proceed in the cytoplasm without the PG precursor being degraded. Although NagA GlcNAc-6-P deacetylase is not, strictly speaking, a recycling enzyme since its substrate is not unique to recycling, it is nevertheless essential for recycling of the cell wall amino sugars. Deletion of one of these cytoplasmic enzymes can lead to a huge accumulation of the substrate for that enzyme in the cytoplasm. For instance, mutants lacking AmpD or NagZ contain at least 100-fold more of their substrate in the cytoplasm (6, 24).

The recycling process begins with degradation of PG by lytic transglycosylases and endopeptidases to release into the periplasm the anhydro-muropeptides, GlcNAc-anhMurNAc-tri-, -tetra-, and -pentapeptides (47). GlcNAc-anhMurNAc-tetrapeptide is the principal product. The pathways for recycling the principal degradation product and the individual PG amino acids are illustrated in Fig. 1. The anhydro-muropeptides are taken up by the cell via the AmpG permease, which is specific for compounds containing GlcNAc-anhMurNAc (7). Once in the cytoplasm, the key enzymes for recycling the murein tripeptide are AmpD, LdcA, and Mpl. AmpD, an anhMurNAc-l-Ala amidase, releases the peptides from the anhydro-muropeptides (21, 25). LdcA, an ld-carboxypeptidase, cleaves d-Ala from the tetrapeptide (49). Mpl, the murein peptide ligase, then links the murein tripeptide to UDP-MurNAc to form UDP-MurNAc-tripeptide, thus returning the peptide to the biosynthetic pathway for PG synthesis (35). Mpl can utilize murein tri-, tetra-, and pentapeptides (20). Thus, LdcA activity is important since in its absence tetrapeptide becomes incorporated into PG, leading to abnormal morphology and eventual lysis in late log phase (49).

Fig. 1Pathways for recycling murein tripeptide and the individual amino acids. Reprinted from reference 44 with permission.

The alternative pathway to utilize the PG amino acids is to break down the tripeptide to the individual amino acids. This is accomplished by the actions of MpaA, a γ-d-Glu-Dap amidase, to produce Dap plus l-Ala-d-Glu; of YcjG, an l-Ala-d/l-Glu epimerase, to convert l-Ala-d-Glu to l-Ala-l-Glu; and of a peptidase such as PepD to cleave l-Ala-l-Glu (46, 51). Even though mpaA and ycjG are expressed during growth, an estimated 90% of the murein tripeptide is directly utilized by Mpl (35).

The pathway for recycling the amino sugars is shown in Fig. 2. GlcNAc-anhMurNAc is released in the cytoplasm either after the GlcNAc-anhMurNAc-peptides taken up via AmpG (7) are cleaved by AmpD (25) or after importation by AmpG into the cytoplasm from the periplasm (7). NagZ, a β-N-acetylglucosaminidase, releases the two amino sugars from the disaccharide GlcNAc-anhMurNAc (6). GlcNAc is then phosphorylated by NagK, a kinase with high specificity for GlcNAc, thus making it available in the normal pathway for amino sugars (52). Most surprising was the finding that anhMurNAc was also converted to GlcNAc-6-P (40). This required two enzymes (54). AnmK kinase breaks the 1,6 anhydro ring present in anhMurNAc and simultaneously phosphorylates the product to produce MurNAc-6-P (54). Then MurQ etherase cleaves the ether bond of MurNAc-6-P, yielding GlcNAc-6-P and d-lactic acid (26, 53). As indicated in Fig. 2, the pathways for uptake and/or utilization of GlcNAc-6-P are well established. Thus, GlcNAc-6-P is readily converted to fructose-6-P or to UDP-GlcNAc and UDP-MurNAc-pentapeptide (32, 33, 34, 57).

Fig. 2Pathway for recycling peptidoglycan aminosugars. Reprinted from reference 44 with permission.

It has been known for 25 years that E. coli can grow on MurNAc as the sole source of carbon and energy (45). Recently it has been demonstrated why E. coli can grow on MurNAc (9). This required a special phosphotransferase system (MurP) as well as MurQ, NagA, and NagB to convert MurNAc to fructose-6-P (Fig. 2) (9). MurP was also found to be required for uptake of anhMurNAc (53). MurQ was shown to be the only MurNAc-6-P etherase in E. coli (53), and hence the pathway for utilization of MurNAc and that for recycling anhMurNAc merge at MurNAc-6-P. Conversion of MurNAc-6-P to GlcNAc-6-P gives the cell a readily metabolizable substrate.

β-LACTAMASE INDUCTION, AN UNFINISHED STORY

Many gram-negative bacteria possess an inducible ampC β-lactamase system that responds to the presence of certain β-lactam antibiotics by producing β-lactamase (38). The inducible system usually consists of an ampC structural gene and an immediately upstream divergently transcribed ampR regulator gene (38). E. coli lacks ampR and hence does not have an inducible β-lactamase (22). The studies cited above were done in E. coli carrying the ampC β-lactamase gene and the ampR gene from Citrobacter freundii on a plasmid.

AmpG and AmpD, both required for recycling, were initially discovered because of their role in ampC β-lactamase induction and hence carry the Amp name (28, 29, 30). AmpG was first identified in Enterobacter cloacae as a gene essential for induction of ampC β-lactamase (28). It is now known that AmpG is a permease required to import cell wall degradation products and that one or more of the recycling intermediates function to induce ampC β-lactamase (24). AmpD, initially described as a negative regulator of β-lactamase, is itself an enzyme that cleaves the anhMurNAc-l-Ala bond (21, 25). The ampD deletion mutant accumulated large amounts of anhMurNAc-tripeptide (24). In a cell-free system to measure ampC messenger synthesis by primer extension, it was shown that AmpR is an activator, that UDP-MurNAc-pentapeptide can repress AmpR, and that the amount of anhMurNAc-tripeptide that accumulates in the ampD mutant can overcome the repression (23).

However, when β-lactamase was induced by the β-lactam antibiotic cefoxitin, there was no accumulation of anhMurNAc-tripeptide (24); hence, anhMurNAc-tripeptide cannot be the true inducer of cells exposed to certain β-lactam antibiotics. Dietz et al. (13), studying E. coli carrying the ampC and ampR genes from Enterobacter cloacae on a plasmid, showed that large increases in the amount of GlcNAc-anhMurNAc-pentapeptide present in the periplasm of ampD ampG cells occurred in cells treated with cefoxitin or imipenem, β-lactams that induce large amounts of β-lactamase, but not in that of ampD ampG cells treated with piperacillin, sulbactam, or cefotaxime, β-lactams that induce little β-lactamase. This striking result certainly implicated a pentapeptide-containing compound in the induction process. However, relatively little anhMurNAc-pentapeptide accumulated in the earlier experiment, whereas a significant amount of a smaller molecular weight compound did accumulate (24). This compound was recently identified as free murein pentapeptide (T. Uehara, K. Suefuji, and J. T. Park, unpublished data). The amount accumulated seems far too little to compete with the large amounts of UDP-MurNAc-pentapeptide present in the cytoplasm. However, it may be that UDP-MurNAc-pentapeptide is not directly involved. It has been reported by Sun et al. (48) that an outer membrane-permeable strain of E. coli, envA, carrying the C. freundii ampC ampR genes on a plasmid, is moderately inducible by moenomycin, vancomycin, fosfomycin, and cycloserine and is highly inducible by ramoplanin as well as by cefoxitin. A possible common link is an effect on the concentration or availability of lipid II. Ramoplanin binds lipid II (14) and would reduce the amount of free lipid II. Also in support of this notion is their observation that a temperature-sensitive murG mutant expresses high levels of β-lactamase at the restrictive temperature (48). MurG is the enzyme that catalyzes the last step in the synthesis of lipid II. Preliminary experiments indicate that a crude preparation of lipid II from 1 ml of E. coli culture prevents production of β-lactamase mRNA in a cell-free system, but possible inducers have not been adequately tested (K. Suefuji, T. Uehara, and J. T. Park, unpublished data). Thus, β-lactamase induction remains an unfinished story.

The induction of β-lactamase by recycling intermediates caused speculation that an intermediate might sense changes in the condition of the murein sacculus and induce expression of wall-related enzymes (24, 42). However, no such effect has been shown in E. coli, possibly because E. coli lacks an AmpR regulator protein. It has recently been shown that in Pseudomonas aeruginosa, AmpR serves as a global transcriptional factor (27). From the observation that P. aeruginosa biofilms exposed to imipenem increase alginate production as well as expression of β-lactamase (1), one could argue that, when treated with β-lactam antibiotics, P. aeruginosa compensates for a weakened wall by responding with alginate production to support the wall.

Another unanswered, interesting question is why expression of ampC β-lactamase is inducible rather than constitutively high. It may be because AmpC overexpression reduces growth and virulent phenotype (36). The mechanism remains to be explained.

PREVALENCE OF THE PG RECYCLING PATHWAYS IN OTHER BACTERIA

Nearly all of the studies of PG recycling have been carried out in E. coli. However, judging by the presence of orthologs of all the genes ampG, ampD, mpl, nagZ, nagK, anmK, and murQ, Salmonella, Shigella, Yersinia, Haemophilus, and Vibrio spp. may recycle both the murein tripeptide and the amino sugars. Curiously, Pseudomonas, Neisseria, and Bordetella spp. appear able to utilize the murein tripeptide but not the amino sugars, since even though they possess an anhMurNAc kinase ortholog, they lack a potential MurQ etherase.

WHY MAY MAMMALS HAVE AN AmpG PERMEASE?

An even more surprising result of the ortholog search is that mammals contain an ortholog of AmpG permease with an identity of about 27% over the entire amino acid sequence. This may be relevant to the process of innate immunity. Nod1 and Nod2 are cytoplasmic proteins in mammals which interact with PG peptides, some of which contain MurNAc or anhMurNAc and Dap, indicating they are products of lysozyme-like enzymes or lytic transglycosylases. The interaction activates a downstream cascade resulting in the innate immune response (5). It has been a puzzle how the PG peptides enter the cytoplasm of mammals. If the AmpG ortholog is active and transports GlcNAc-MurNAc-peptides and/or GlcNAc-anhMurNAc-peptides, it would explain how PG fragments enter the cytoplasm to be detected by Nod1 and Nod2.

CONCLUSIONS

PG recycling is an active process that allows E. coli to reuse over 60% of the PG of the sidewall each generation. On the basis of the presence of orthologs, many other bacteria may recycle their PG amino acids and amino sugars. Increased concentrations of a PG recycling intermediate are required for an inducible β-lactamase system to be expressed. However, the exact nature of the inducer and corepressor molecules remains to be established.