Ngg triggering first the synthesis of the intermediate NAGG and
AsnO its subsequent conversion into NAGGN. These two pro-
is also obvious that NAGGN is produced within a taxonomically
and physiologically diverse set of bacterial species. In particular,
teins represent a unique enzymatic machinery mediating non- the presence of asnO–ngg cluster on the genome of several ma-
ribosomal peptide synthesis in bacteria, which differs from the rine bacteria and in bacteria subjected to osmolarity fluctuations
nonribosomal peptide synthetases, the poly-γ-glutamate, cyano- in their environment indicate that NAGGN could participate in
phycin, and glutathione synthetases, the D-alanine:D-alanine li- the responsiveness to seawater salinity and osmolarity variation.
gase, the L-amino acid α-ligase, and others (24).
Our results (Fig. 4A) show that the growth of an asnO mutant
that does not produce any NAGGN is much more affected than
the growth of the WT strain in a medium of elevated osmolarity,
and thus offer definitive evidence of the efficiency of the di-
peptide for osmoprotection of the cells. Interestingly, the asnO–
ngg cluster of S. meliloti is located within a region of the pSymb
replicon that has been shown recently to be important for osmo-
protection, and hence for the survival of the free-living form in
the rhizosphere (31). Our analysis with an asnO transcriptional
fusion revealed that, in the absence of exogenous osmoprotective
compound, an increasing NaCl concentration triggers an in-
crease in the expression of the asnO–ngg operon (Fig. 4C). It
is also noteworthy that the presence of exogenous compatible
solute, such as GB, has a reverse effect on this induction (Fig.
4D). Thus, it is obvious that the cells sensitively adjust the ex-
pression of the asnO and ngg genes with a direct consequence on
the NAGGN content. It has already been assumed that the
biosynthesis of organic compatible solute is energetically more
costly than the uptake of osmolytes from the environment when
available (32).
All the genomes carrying the asnO–ngg cluster encode a con-
served putative peptidase, with the exception of Mesorhizobium
sp., M. avium, M. vanbaalenii, M. gilvum, M. smegmatis, and M.
sp. MCS. Two different organizations are observed: (i) the gene
encoding the peptidase is located immediately downstream ngg
for 26 of the 40 asnO–ngg clusters, such as in P. aeruginosa (ORF
PA3461), or (ii) this gene is found elsewhere on the genome in
eight cases, like in S. meliloti (ORF SMb20466). This peptidase
belongs to the M42 peptidase family that contains aminopepti-
dases hydrolyzing acylated N-terminal residues (33). Despites the
fact that its role still remains to be established, one can suggest that
such peptidase could play a role in balancing NAGGN pool during
adaptation to osmotic fluctuations. In addition, particularly during
an osmotic down-shock, a direct release of the dipeptide cannot be
excluded. The presence of genes encoding a putative dipeptide
ABC transporter located directly upstream of asnO might suggest
such possibility or an efflux of hydrolysates.
Another attractive and intriguing feature about the asnO–ngg
cluster comes from the work of Berges and collaborators (11).
The authors have shown that a mutation in the asnO gene im-
paired the activity of FixT, an antikinase that inhibits the FixL–
FixJ two-component system that controls the expression of ni-
trogen fixation genes in bacteroid, the symbiotic form of S. meli-
loti. It is thus tempting to argue that the NAGGN biosynthetic
pathway could play an important role inside the nitrogen-fixing
nodules. These data, together with the presence of an asnO–ngg
cluster on the genome of pathogenic or symbiotic bacteria, offer
insights into the relationships between osmoadaptation and host
interactions. In view of the variety of stresses encountered by
pathogenic bacteria during the course of infection, a number of
osmoprotective compounds have already been linked to the vir-
ulence potential of certain pathogens (34). Therefore, our study
raises questions about the effectiveness of NAGGN as a latent
virulence factor in pathogenic bacteria. AsnO and Ngg proteins,
which are conserved in animal and human pathogens (M. avium,
P. aeruginosa, Pseudomonas mendocina, and O. anthropi), but
absent in eukaryotes, could be consequently attractive targets for
antibacterial drugs.
The dipeptide formation is mediated by Ngg, which appears to
be a bifunctional protein also catalyzing an acetylation reaction.
Peptide synthetases with modular organization have been de-
scribed previously. However, conversely to Ngg and excepting
some nonribosomal peptides (25), they catalyze only peptide-
forming reactions. For example, CyaH incorporates an Asp or
Arg residue into a cyanophycin primer using two distinct ATP-
dependent ligase activities (19), and the synthesis of the tripeptide
glutathione (γ-glutamyl-cysteinylglycine) in S. agalactiae is medi-
ated by a bifunctional enzyme, GshAB, that contains the γ-glu-
tamylcysteine and glutathione synthetase activities usually found
on two distinct proteins (26). By its N-terminal domain, the Ngg
protein belongs to the enormous family of GCN5-related N-ace-
tyltransferases. In bacteria, these enzymes participate in impor-
tant metabolic pathways such as the synthesis of spermidine and
mycothiol, or in the extension mechanism of pepdidoglycan in
Gram-positive bacteria (15, 27). An example of a bifunctional
enzyme containing a GNAT domain is GlmU, which catalyzes the
acetylation of glucosamine-1-phophate and its subsequent uridy-
lylation to give UDP-N-acetylglucosamine (28). Multiple repre-
sentatives of this family are present on any sequenced genome,
but most of them have not been functionally characterized and
their substrates are still unknown. Within the S. meliloti genome,
22 of the 38 ORFs annotated as encoding acetyltransferases be-
long to the GNAT family, two have been ascribed as acetylating
ribosomal proteins, and one as a GlmU homologue. The dem-
onstration that Ngg is involved in the biosynthesis of NAGG is
a unique functional characterization of a GNAT protein in this
symbiotic bacterium. Our results give a molecular support to the
previous in vitro experiments that have proposed the formation of
NAGGN with NAGG as a potential intermediate (9). However,
conversely to these preliminary results, the bioinformatic analysis
of the Ngg sequence indicated that NAGG formation can occur
directly from two glutamines and not from one N-acetylglutamine
and one glutamine. A possible explanation for this discrepancy is
the absence of AcCoA in the in vitro reaction mixture, and thus
only the Ngg activity making a peptide bond between N-acetyl-
glutamine and glutamine could be seen. Nevertheless, additional
biochemical work is still needed to establish (i) the sequence of
the two reactions catalyzed in vivo by Ngg and (ii) if N-acetyl-
glutamine can directly be used in vivo as a substrate.
The asnO gene is required for the final step of NAGGN syn-
thesis, i.e., the addition of a NH2 group to the NAGG moiety thus
creating a C-terminal amide group. As a member of the PurF
family, AsnO most likely catalyzes the removal of the NH2 group
from glutamine via its N-terminal domain, then transfers it to
NAGG within its C-terminal domain. This is an example of an
enzyme similar to the E. coli AsnB that is not involved in aspar-
agine synthesis (29, 30). Sequence comparisons indicate that the
proteins closely related to the S. meliloti AsnO group into the
amidotransferase AsnB family, but that they are distinct from
previously identified members of this family (Fig. S4). Such dis-
tribution supports the idea that AsnO-like proteins have their own
catalytic properties. Consequently, the presence of AsnO homo-
logues in bacteria that do not encode a Ngg homologue, such as in
Frankia alni or in Methylobacterium extorquens, remains in-
triguing, and its biological significance is still far from clear.
The high degree of sequence identity among the AsnO and
Ngg proteins and the preservation of the asnO–ngg organization
in various bacteria showed that the NAGGN biosynthetic path-
way is evolutionally well conserved in the bacterial kingdom. It
Finally, the characterization of the asnO–ngg cluster offers a set
of biosynthetic genes for metabolic engineering technology, to in-
stall the production of NAGGN in bacteria or agriculturally im-
portant crop plants to improve their tolerance to osmotic stress.
12656
|
Sagot et al.