Arch Microbiol (2000) 173 : 170177
Digital Object Identifier (DOI) 10.1007/s002039900127
MINI-REVIEW
Caroline Blumer Dieter Haas
Mechanism, regulation,
and ecological role of bacterial cyanide biosynthesis
Received: 30 August 1999 / Revised: 12 November 1999 / Accepted: 15 November 1999 / Published online: 11 January 2000
Springer-Verlag 2000
Abstract A few bacterial species are known to produce
and excrete hydrogen cyanide (HCN), a potent inhibitor
of cytochrome c oxidase and several other metalloen-
zymes. In the producer strains, HCN does not appear to
have a role in primary metabolism and is generally con-
sidered a secondary metabolite. HCN synthase of pro-
teobacteria (especially fluorescent pseudomonads) is a
membrane-bound flavoenzyme that oxidizes glycine, pro-
ducing HCN and CO2. The hcnABC structural genes of
Pseudomonas fluorescens and P. aeruginosa have se-
quence similarities with genes encoding various amino
acid dehydrogenases/oxidases, in particular with nopaline
oxidase of Agrobacterium tumefaciens. Induction of the
hcn genes of P. fluorescens by oxygen limitation requires
the FNR-like transcriptional regulator ANR, an ANR re-
cognition sequence in the 40 region of the hcn promoter,
and nonlimiting amounts of iron. In addition, expression
of the hcn genes depends on a regulatory cascade initiated
by the GacS/GacA (global control) two-component sys-
tem. This regulation, which is typical of secondary metab-
olism, manifests itself during the transition from exponen-
tial to stationary growth phase. Cyanide produced by P.
fluorescens strain CHA0 has an ecological role in that this
metabolite accounts for part of the biocontrol capacity of
strain CHA0, which suppresses fungal diseases on plant
roots. Cyanide can also be a ligand of hydrogenases in
some anaerobic bacteria that have not been described as
cyanogenic. However, in this case, as well as in other sit-
uations, the physiological function of cyanide is un-
known.
Key words Hydrogen cyanide Hydrogen cyanide
synthase Opine oxidase Anaerobic control Iron
regulation Pseudomonas fluorescens Pseudomonas
C. Blumer D. Haas ()
Laboratoire de Biologie Microbienne, Universit de Lausanne,
CH-1015 Lausanne, Switzerland
e-mail: Dieter.Haas@lbm.unil.ch,
Tel.: +41-21-6925631, Fax: +41-21-692-5635
aeruginosa Chromobacterium violaceum Biocontrol
Hydrogenase
Introduction
Some bacteria produce a potent poison: cyanide. Yet, there
seem to be no reports describing acute or chronic toxicity
caused by bacterial cyanide in animals or plants. One rea-
son for this may be the fact that cyanide production (cyano-
genesis) is very tightly regulated in bacteria, such that the
local cyanide concentrations usually are below 1 mM and
can be tolerated by many living cells. However, it is pos-
sible that host cells when massively infected by a cyano-
genic pathogen (such as Pseudomonas aeruginosa) may
suffer from deleterious effects caused by cyanide (Gold-
farb and Margraf 1967). The main purpose of this review
is to discuss some recent molecular studies on cyanide
production by fluorescent pseudomonads in the light of
older, well established facts (reviewed by Vennesland et
al. 1981a; Knowles and Bunch 1986; Knowles 1988).
Under physiological conditions at pH 7, cyanide is
mostly present as hydrogen cyanide (HCN), according to
its pKa of 9.3. HCN, being volatile and less dense than air,
can rapidly diffuse into the environment. In this review,
we do not distinguish between cyanide and HCN, and we
will focus on cyanide biosynthesis in bacteria. Cyanogen-
esis has also been described in fungi, algae and plants (re-
viewed by Vennesland et al. 1981a; Knowles and Bunch
1986; Jones 1998). Whereas no specific role has been as-
cribed to cyanide of fungal or algal origin, cyanogenesis
in a wide range of plants is believed to constitute a chem-
ical defense against herbivores and pathogens. Plants use
cyanogenic glucosides (e.g. amygdalin) as precursors of
cyanide (Jones 1998); this hydrolytic pathway differs fun-
damentally from the bacterial cyanide biosynthetic path-
way discussed here. Cyanide detoxification mechanisms
are widespread in nature and involve enzymatic degrada-
tion or conjugation with acceptor compounds such as cys-
teine, thiosulfate or -keto acids (Evered and Harnett
1988; Knowles 1988; Raybuck 1992; Kunz et al. 1998).

Fig. 1 HCN synthase reaction
and related enzymatic reactions
according to: a Laville et al.
(1998), b Nishiya and Imanaka
(1998), c Zanker et al. (1994),
and d Vennesland et al.
(1981b)
Bacterial cyanogenesis appears to be essentially re-
stricted to the proteobacteria Chromobacterium violaceum
and to fluorescent pseudomonads (including many strains
of P. aeruginosa and P. fluorescens, and some isolates of
P. aureofaciens and P. chlororaphis) as well as certain
cyanobacteria (Anacystis nidulans, Nostoc muscorum and
Plectonema boryanum) (Castric 1981; Vennesland et al.
1981b; Knowles and Bunch 1986). Recently, some strains
of Rhizobium leguminosarum have been reported to pro-
duce HCN as free-living bacteria (Antoun et al. 1998).
171
Metabolic pathway
and physiology of bacterial cyanogenesis
In vivo studies have shown that glycine is the immediate
metabolic precursor of cyanide in proteobacteria. HCN
and CO2 are formed stoichiometrically by oxidative de-
carboxylation of glycine (Michaels et al. 1965; Wissing
1974; Castric 1977). In experiments using radiolabeled
[1-14C]glycine or [2-14C]glycine as the substrate, cyanide
was found to be derived from the methylene carbon of
glycine and CO2 from the carboxyl group of glycine in C.
violaceum (Bunch and Knowles 1982), P. fluorescens and
P. aeruginosa (Askeland and Morrison 1983). The CN
bond is retained during the reaction (Brysk et al. 1969),
excluding possible transamination or deamination reac-
172
tions. In P. aeruginosa, threonine can be metabolized to
glycine and also serve as a precursor for cyanide, albeit
less effectively (Castric 1977). In vitro, the enzyme com-
plex converting glycine to HCN and CO2, HCN synthase,
appears to be membrane-associated, both in a Pseudo-
monas species and C. violaceum, and can be solubilized
by detergents (Wissing 1975; Wissing and Andersen 1981;
Bunch and Knowles 1982). HCN synthase is very labile
and sensitive to oxygen, although glycine protects the
Pseudomonas and Chromobacterium enzymes somewhat
from oxygen toxicity (Castric 1981; Wissing and Ander-
sen 1981; Bunch and Knowles 1982). Because of these
stability problems, HCN synthase has been purified only
partially from a Pseudomonas species (Wissing and An-
dersen 1981). As a consequence, the subunit composition
and cofactor requirements of the enzyme are unknown.
In crude extracts, HCN synthase of Pseudomonas sp.,
P. aeruginosa and C. violaceum can use artificial electron
acceptors, e.g. phenazine methosulfate (PMS), to oxidize
glycine (Wissing 1975; Castric 1981; Bunch and Knowles
1982). Whereas in vivo the natural electron acceptor is
oxygen, in vitro the acceptor for the four electrons pro-
duced by oxidative decarboxylation of glycine has yet to
be identified. It is probable that components of the respi-
ratory system are involved (Castric 1994). However, oxy-
gen is not strictly required for cyanogenesis by C. vio-
laceum in which fumarate can be the terminal electron ac-
ceptor (Nazly et al. 1981). Importantly, HCN synthase of
Pseudomonas sp. is strongly inhibited in whole cells by
pyrrolnitrin, an inhibitor of flavine enzymes, and by the
membrane-permeable metal chelator o-phenanthroline
(Wissing 1974). In vitro, the addition of FAD stimulates
HCN synthase activity (Wissing 1975).
Fig. 2 Alignment of the ADP-binding motif in the HcnB and
HcnC subunits of HCN synthase from Pseudomonas fluorescens
(GenBank accession no. AF053760) and related proteins. OoxA
and OoxB, subunits of octopine oxidase from Agrobacterium
tumefaciens (GenBank accession no. Z30328); NoxA and NoxB,
subunits of nopaline oxidase from A. tumefaciens (GenBank ac-
cession no. Z30316); SoxA and SoxB, subunits of sarcosine oxi-
dase from Corynebacterium sp. (GenBank accession no. U23955);
GoxB, glycine oxidase from Bacillus subtilis (GenBank accession
no. Z99110); DadA, hypothetical D-amino acid dehydrogenase
from P. aeruginosa (GenBank accession no. L48934). Identical
residues are boxed; conserved glycine residues (G) are indicated in
boldface and the other conserved residues that define the motif de-
scribed by Wierenga et al. (1986) are marked by asterisks; the
-fold is shown above the sequence
Based on these findings, Wissing (1974) has proposed
a model (Fig. 1a) according to which glycine is oxidized
in two steps, with iminoacetate as an unstable intermedi-
ate. HCN synthase is postulated to be a membrane-bound
flavoprotein (Wissing 1974; Wissing and Andersen 1981).
Several other enzymatic mechanisms have been proposed
(Knowles and Bunch 1986). However, none of these are
supported by sequence data (discussed below). Therefore,
the systematic name glycine dehydrogenase (cyanide-
forming; EC 1.4.99.-) is proposed for HCN synthase.
The structural genes hcnABC
for HCN biosynthesis have similarities
with genes encoding amino acid oxidases
The HCN biosynthetic genes have recently been identi-
fied in P. fluorescens strain CHA0 (Laville et al. 1998)
and P. aeruginosa PAO (G. Pessi and D. Haas, unpub-
lished data). A cluster of three genes, hcnABC, which pre-
sumably form an operon, encodes HCN synthase. When
expressed from the T7 promoter, the hcnABC genes of
P. fluorescens are sufficient to render Escherichia coli
cyanogenic (Laville et al. 1998). However, it is important
to use freshly transformed E. coli recombinant cells, as
they poorly tolerate the presence of the hcn genes. The de-
duced amino acid sequences of the three HCN synthase
subunits of P. fluorescens have similarities with known
dehydrogenases or oxidases. The HcnB and HcnC poly-
peptides having predicted molecular masses of 50,647 Da
and 45,334 Da, respectively, each contain a typical FAD-
or NAD-binding motif, the ADP-binding -fold
(Wierenga et al. 1986), in their N-terminal parts (Fig. 2).
HcnB is similar to the OoxA subunit of octopine oxidase
(30% identity) and to the NoxA subunit of nopaline oxi-
dase (30% identity) from Agrobacterium tumefaciens
(Zanker et al. 1994), as well as to the SoxA subunit of sar-
cosine (N-methyl glycine) oxidase (32% identity) from
Corynebacterium sp. (Chlumsky et al. 1995). HcnC has
similarities with the corresponding subunits from the
same organisms, OoxB (24% identity), NoxB (23% iden-
tity), and SoxB (23% identity). HcnC also resembles the
newly described flavoenzyme glycine oxidase (GoxB;
26% identity) of Bacillus subtilis (Nishiya and Imanaka
1998). The reaction catalyzed by glycine oxidase is
shown in Fig. 1b. The DadA subunit of a putative D-amino
acid oxidase from P. aeruginosa (GenBank accession no.

Fig. 3 Comparison of the amino acid sequences of the hypotheti-
cal NoxC and OoxC subunits of nopaline oxidase from A. tumefa-
ciens and octopine oxidase from Rhizobium meliloti, respectively,
with HcnA subunit of HCN synthase from P. fluorescens. Con-
served amino acid residues are indicated in boldface. The four cys-
teine residues that are believed to bind an [2Fe2S] cluster are
marked by triangles beneath the sequence
L48934) also shows similarity (27% identity). An align-
ment of the N-terminal parts of the above proteins, all
of which contain an ADP-binding motif, is shown in
Fig. 2.
The small HcnA polypeptide (11,525 Da) has 31%
identity with the HoxU subunit of hydrogenase from An-
abaena variabilis (Schmitz et al. 1995) and 34% identity
with the subunit of formate dehydrogenase (FdhA) from
Moorella thermoacetica (GenBank accession no. U73807).
It contains a cluster of four cysteines (Fig. 3), which
matches the iron-sulfur binding signature of [2Fe2S]
ferredoxins (documented in the PROSITE database http://
www.expasy.ch/cgi-bin/get-prodoc-entry?/PDOC00175).
Ferredoxins are iron-sulfur proteins that mediate electron
transfer in a wide range of metabolic reactions. HoxU and
FdhA are assumed to bind [2Fe-2S] in the regions homo-
logous to HcnA (Schmitz et al. 1995).
Interestingly, the nucleotide sequence of the hcnA gene
shows 56.2% identity with the nucleotide sequence of a
242 bp region between ooxB and ooxA, the genes encod-
ing a putative octopine oxidase, from Rhizobium meliloti
(Genbank accession no. U66830). Sequence analysis of
this region reveals a small ORF. A similar, somewhat
longer ORF (294 bp) is found in the region between noxB
and noxA, the genes coding for nopaline oxidase from A.
tumefaciens (GenBank accession no. Z30316). The de-
duced amino acid sequences of these ORFs, which we
designate ooxC and noxC, respectively, are similar to that
of HcnA, particularly in the region of the clustered cys-
teines, which are believed to bind a [2Fe-2S] center (Fig.
3). Expression of the isolated noxA and noxB genes from
separate plasmids failed to produce nopaline oxidase ac-
tivity, whereas a contiguous noxB (noxC) noxA construct
did lead to activity (Zanker et al. 1994). This result can
now be explained by assuming that NoxC is part of a
functional nopaline oxidase; the reaction is shown in Fig.
1c. In the original study of Zanker et al. (1994), the NoxC
peptide was missed, probably because of its small size.
The ooxB-ooxA intergenic region of A. tumefaciens may
also contain an additional ORF, but this is interrupted by
an internal stop codon (data not shown). Unfortunately,
the mechanism of the opine oxidase reaction and the co-
173
factors involved remain obscure, as attempts to purify the
enzyme have not been successful.
Taken together, the sequence comparisons suggest that
HCN synthase basically is an amino acid dehydroge-
nase/oxidase, which supports the hypothetical pathway
proposed in Fig. 1a. Glycine oxidase of B. subtilis and re-
lated D-amino acid oxidases of various organisms are
monosubunit flavoproteins catalyzing oxidation of the
substrate amino acids to the corresponding imino acids,
which are hydrolyzed immediately to the -keto acids and
ammonia (Fig. 1b; Hafner and Wellner 1979; Nishiya and
Imanaka 1998). Both glycine oxidase (Fig. 1b) and opine
oxidase (Fig. 1c) cleave CN bonds. By contrast, in the
HCN synthase reaction (Fig. 1a), the postulated enzyme-
bound intermediate iminoacetate is cleaved at the CC
bond. The biochemical action underlying this cleavage is
not clear. However, it is known from work on cyanobac-
teria that amino acid oxidases, when coupled to peroxi-
dase, can produce HCN (Vennesland et al. 1981b). For in-
stance, Anacystis nidulans uses L-histidine as a precursor
of HCN, in a reaction catalyzed by an L-amino acid oxi-
dase. In the absence of peroxidase, the imino derivative of
histidine is formed and rapidly hydrolyzed to give imida-
zole pyruvate. When peroxidase is added, some of the
imino acid is not hydrolyzed but converted to HCN, CO2
and imidazole aldehyde, for reasons that are not entirely
evident (Fig. 1d). L-Histidine is not the only amino acid
that yields HCN in this coupled system and peroxidase
can be replaced by inorganic oxidants (Vennesland et al.
1981b).
It may also be that cleavage of the CC bond in the
HCN synthase reaction could involve a radical mecha-
nism similar to that employed by E. coli pyruvate for-
mate-lyase, which catalyzes the nonoxidative dissimila-
tion of pyruvate to formate and acetyl-CoA (Sawers
1999). In the activated enzyme, a catalytically active thiyl
radical, generated by the transfer of an unpaired electron
from an adjacent glycyl radical in the C-terminus of the
protein, initiates the cleavage of pyruvate. In HcnC, the
amino acid sequences surrounding Cys-148 and Gly-362,
located at the C-terminal part (Laville et al. 1998), show
some similarities with pyruvate formate-lyase and other
glycyl radical enzymes (Sun et al. 1996; Leuthner et al.
1998). However, the resemblance is weak and might be
fortuitous.
The association of HCN synthase with the cytoplasmic
membrane (Wissing and Andersen 1981; Knowles and
Bunch 1986) can be understood in terms of the deduced
amino acid sequences of HcnB and HcnC, which both
174
Fig. 4 Promoter region of the
hcn genes of P. fluorescens
CHA0 and activation by ANR.
The ANR box (40 region) and
the 10 region are underlined;
+1 transcription start site. This
model is inspired by the FNR
models of Unden and Schi-
rawski (1997) and Kiley and
Beinert (1999). The presence
of an FeS cluster in ANR has
not been demonstrated experi-
mentally, but is inferred from
the extensive structural and
functional similarities between
FNR and ANR (Laville et al.
1998; Hjberg et al. 1999).
Lack of iron and/or the pres-
ence of oxygen can inactivate
ANR. ANRapo, inactive apo-
protein without FeS cluster
contain two putative transmembrane segments (Laville et
al. 1998). The topology of HCN synthase may be such
that cyanide, a good nucleophile and strong chelator of
certain metal ions (Ni2+, Cu2+, Co2+, etc.), will be released
into the periplasm rather than into the cytoplasm where
this product would have undesirable toxic effects.
Regulation of hydrogen cyanide production
in pseudomonads
Glycine, the metabolic precursor of HCN, stimulates
cyanide formation in P. aeruginosa (Castric 1977) and P.
fluorescens CHA0 (Voisard et al. 1989). This effect pre-
sumably operates at a posttranslational level, since the ex-
pression of a translational hcnA-lacZ fusion of P. fluo-
rescens CHA0 is not influenced by the addition of glycine
to a defined medium (our unpublished results). Added
glycine may directly act as a substrate and/or stabilize
HCN synthase (Castric et al. 1981).
HCN is produced maximally by pseudomonads during
the transition from the exponential to the stationary phase;
this idiophase is followed by a rapid loss of HCN synthase
activity (Castric et al. 1979, 1981; Askeland and Morrison
1983). In addition, cyanogenesis requires oxygen, but in
limiting amounts, i.e. microaerophilic conditions (Castric
1983; Castric et al. 1981; Laville et al. 1998). In P. aerugi-
nosa and P. fluorescens, HCN production depends strictly
on the ANR protein (anaerobic regulator of arginine
deiminase and nitrate reductase); anr mutants of both or-
ganisms produce very little HCN (Zimmermann et al.
1991; Laville et al. 1998). ANR belongs to the FNR (fu-
marate and nitrate reductase regulator) family of tran-
scriptional regulators (Spiro 1994). FNR of E. coli is a
dimeric oxygen-sensing protein whose two [4Fe4S]2+
clusters are converted to the [2Fe2S]2+ form in the pres-
ence of oxygen, resulting in the inactivation of the pro-
tein. Strong aeration leads to loss of the iron-sulfur cluster
and formation of the monomeric FNR apoprotein (com-
pare with Fig. 4). The reverse reaction, i.e. assembly of
the iron-sulfur cluster on FNR, needs Fe2+, S2 and en-
zymes such as the NifS homologue (Kiley and Beinert
1999). Below a concentration of 510 M of dissolved
oxygen, FNR can bind to conserved sequences (known as
the FNR box) in promoter regions of target genes. De-
pending on the position of the FNR box relative to the
transcription start site, activation or repression of the tar-
get promoter can occur (Spiro 1994; Becker et al. 1996;
Rhodius and Busby 1998). The ANR proteins of P. aerugi-
nosa and P. fluorescens CHA0 exhibit extensive struc-
tural and functional similarity to FNR (Sawers 1991; Zim-
mermann et al. 1991; Laville et al. 1998). Whereas P.
aeruginosa is a facultative anaerobe and grows with ni-
trate as the electron acceptor, P. fluorescens CHA0 is an
obligate aerobe and unable to denitrify. However, strain
CHA0 grows well under microaerophilic conditions, e.g.
at 2 M O2, which is sufficiently low to activate ANR
(Hjberg et al. 1999).
The promoter of the hcnABC genes of P. fluorescens
CHA0 (Laville et al. 1998) contains an FNR/ANR box in
the 40 region, which is characteristic of promoters acti-
vated by FNR or ANR (Fig. 4). Studies using a transla-
tional hcnA-lacZ fusion driven from the native promoter
with or without the ANR box have shown that induction
of hcn expression by oxygen limitation strongly depends
on ANR and the ANR box (Blumer et al. 1999). Hence,
HCN synthase is controlled by the oxygen tension at two
levels: at the level of transcription via ANR and at the
level of enzyme activity.
Iron has a stimulatory effect on cyanogenesis in C. vio-
laceum, P. aeruginosa and P. fluorescens (Castric 1975;
Rodgers and Knowles 1978; Askeland and Morrison
1983; Voisard et al. 1989). In P. fluorescens CHA0, this
response to iron is mediated by ANR and depends on the

presence of the ANR box in the hcn promoter. When the
natural, strongly ANR-regulated hcn promoter is replaced
by an artificial promoter that weakly responds to oxygen
limitation, iron regulation is lost (Blumer and Haas, un-
published data). Iron limitation might restrain the assem-
bly of Fe-S clusters on the ANR apoprotein, favoring in-
activation of ANR even under oxygen-limiting conditions
(Fig. 4).
Cyanogenesis in P. fluorescens and P. aeruginosa is
controlled by a second regulatory protein: the global acti-
vator GacA. GacA is the response regulator of the con-
served, two-component regulatory system GacS (formerly
LemA)/GacA, which has a decisive role in the control of
a variety of extracellular products in many gram-negative
bacteria (Laville et al. 1992; Reimmann et al. 1997). In P.
aeruginosa, GacA positively controls the expression of
the cell density signal N-butyryl-homoserine lactone,
which activates the LuxR-like transcriptional activator
RhlR and thereby induces several genes coding for exo-
products including the hcnABC genes (Winson et al.
1995; Reimmann et al. 1997). This induction requires a
lux box (i.e. a putative RhlR recognition site) at about 70
in the hcnA promoter (G. Pessi and D. Haas, unpublished
data). Interestingly, the hcnA promoter of P. fluorescens
CHA0 lacks a lux box (Fig. 4) and there is no evidence for
N-acyl-homoserine lactone signal molecules in this organ-
ism. It appears that GacA control of cyanogenesis in
strain CHA0 follows a different and novel regulatory
pathway, acting essentially at a posttranscriptional level
(Blumer et al. 1999).
Cyanide, a typical secondary metabolite
Cyanide production by proteobacteria has several charac-
teristics of secondary metabolism (Castric 1975; Vining
1990):
1. Cyanide has no apparent function in primary metabo-
lism and is optimally produced during growth limita-
tion (by oxygen).
2. The producer organism is tolerant to cyanide.
3. Cyanogenesis has an ecological role and may offer the
producer a selective advantage.
Let us consider tolerance first. Micromolar amounts of
cyanide strongly inhibit cytochrome c oxidase, the termi-
nal component of the respiratory chain in many organ-
isms, as well as several other metalloenzymes, e.g. cata-
lase, peroxidase, superoxide dismutase, nitrate reductase,
nitrite reductase, and nitrogenase (Solomonson 1981). In
cytochrome c oxidase, the CuB2+ site is involved in initial
binding of cyanide (Wilson et al. 1994). P. aeruginosa
displays cyanide-resistant respiration during HCN synthe-
sis, apparently in response to the presence of HCN (Kra-
lik and Castric 1979). A cyanide-insensitive terminal oxi-
dase, CioAB, has been identified in this organism (Cun-
ningham et al. 1997). Therefore, the CioAB oxidase
might allow aerobic respiration under cyanogenic growth
conditions. However, ANR negatively regulates the cioAB
175
genes (Cunningham et al. 1997) and it is therefore diffi-
cult to see how the expression of this terminal oxidase
would be synchronized with cyanogenesis.
Under denitrifying conditions, i.e. in the presence of
nitrate and the absence of oxygen, P. aeruginosa does not
produce cyanide (Castric 1975). By this unexplored
mechanism, a conflict is avoided, in that inhibition of ni-
trate and nitrite reductases by cyanide (Solomonson 1981)
is prevented. P. aeruginosa also has some potential to
detoxify external cyanide. Under conditions of nitrogen
limitation, this bacterium excretes -ketoglutarate (Von
Tigerstrom and Campbell 1966), a compound that in-
stantly combines with cyanide in a nonenzymatic, nucle-
ophilic addition reaction, producing a less toxic cyanohy-
drin. Kunz et al. (1998) have recently provided evidence
that P. fluorescens NCIMB 11764 can scavenge and uti-
lize cyanide as a nitrogen source by excreting copious
amounts of -ketoglutarate and pyruvate. The resulting
cyanohydrins are then degraded intracellularly, presum-
ably by cyanide oxidase. This reaction regenerates the -
keto acid and produces CO2 and NH3 (Kunz et al. 1998).
In this case, tolerance to cyanide and the ability to utilize
it as a cyanohydrin are coupled. However, the two pro-
cesses need not always occur together.
An ecological role for bacterial cyanogenesis has been
discovered for plant-beneficial strains of P. fluorescens
and P. putida. The root-colonizing P. fluorescens strain
CHA0 protects several plants from fungal root diseases
(Voisard et al. 1994). As shown by mutant studies, HCN
production accounts for a substantial part of the strains
biocontrol capacity, e.g. the suppression of tobacco black
root rot caused by Thielaviopsis basicola (Voisard et al.
1989; Laville et al. 1998). HCN formation by a recombi-
nant strain of P. putida BK8661 seems to be involved in
the suppression of disease symptoms caused by Septoria
tritici and Puccinia recondita f. sp. tritici on wheat seedling
leaves (Flaishman et al. 1996). In vitro, HCN produced by
cultures of P. fluorescens on solid medium inhibits the
growth of several phytopathogenic fungi such as T. basi-
cola or Gaeumannomyces graminis via the gas phase (C.
Keel, personal communication). However, there are no
data available demonstrating the inhibition of microor-
ganisms by HCN in the rhizosphere or in soil. Bakker and
Schippers (1987) have presented a hypothesis (but little
evidence) that some Pseudomonas strains, by producing
cyanide in the rhizosphere, could depress growth of
potato.
In wounds infected by P. aeruginosa, cyanide has been
detected. This has led to the speculation that cyanide may
be a virulence factor of this opportunistic pathogen (Gold-
farb and Margraf 1967). Supporting evidence has recently
been obtained by Gallagher and Manoil (1999), who
found that an hcnC mutant of P. aeruginosa had a strongly
reduced ability to kill the nematode Caenorhabditis ele-
gans.
Cyanide forms stable complexes with transition metal
ions and, therefore, it might be thought that such com-
plexes could help the producer bacteria to mobilize some
of these metal ions, e.g. from clay minerals in soil. The
176
very stable ferro- and ferricyanide complexes (K4Fe(CN)6,
K3Fe(CN)6) do not form under physiological conditions.
However, a culture medium of a Pseudomonas sp. grown
under iron-limiting conditions was found to contain a
mixed iron-cyano complex, dicyano-bis(pyridine-2,6-di-
carbothioato)-ferrate, which constitutes a reversible fer-
rate(II)/ferrate(III) redox system. The redox potential
(E0=+0.400 V) of this complex is comparable to that of
certain cytochromes (Hildebrand et al. 1985). Unfortu-
nately, the physiological function of this complex is un-
known.
Cyanide, a ligand of hydrogenases
Recent studies revealed that cyanide occurs as an intrinsic
ligand of the prosthetic group of both [NiFe]- and [Fe]-
hydrogenases in bacteria oxidizing or evolving hydrogen.
The iron in the active site of the [NiFe]-hydrogenase of
Chromatium vinosum contains two cyanide groups and
one carbon monoxide molecule as ligands (Pierik et al.
1999); the active site Fe center in oxygen-sensitive [Fe]-
hydrogenases of Desulfovibrio vulgaris substrain Hilden-
borough (Pierik et al. 1998) and of D. desulfuricans
(Nicolet et al. 1999) is coordinated by one cyanide and
one carbon monoxide molecule. The origin of the cyanide
ligand in these strictly anaerobic, noncyanogenic organ-
isms is not known. By contrast, cyanogenic bacterial
species of the genera Chromobacterium and Pseudo-
monas have not been reported to possess hydrogenases.
The filamentous cyanobacterium Nostoc muscorum, one
of the few photosynthetic bacteria known to be cyano-
genic (Vennesland et al. 1981b), contains two different
types of [NiFe]-hydrogenases (Axelsson et al. 1999).
Whether these contain cyanide ligands at their active sites
remains to be elucidated.
Concluding remarks
In pseudomonads, cyanogenesis, a process of secondary
metabolism, depends on two regulatory elements, the
transcriptional regulator ANR and the GacS/GacA two-
component system. ANR is shared with primary metabo-
lism. By contrast, as suggested by our recent studies on
hcn gene expression (Blumer et al. 1999), a critical switch
from primary to secondary metabolism seems to be oper-
ated by the GacS/GacA system.
In common with many other secondary metabolites,
cyanide has not yet been assigned a physiological function
in bacterial metabolism, but an ecological role has been
found. Bacterial cyanide can have an advantage for plants:
it helps them in their defense against fungal pathogens.
Acknowledgements Work done in the authors laboratory was
supported by the Swiss National Science Foundation (projects 31-
45896.95 and 31-50522.97), the EU projects BIO4CT960119 and
BIO4CT960027, and the Swiss Priority Program Biotechnology
(project 5002-04502311). We thank Cornelia Reimmann for dis-
cussions.
References
Antoun H, Beauchamp CJ, Goussard N, Chabot R, Lalande R
(1998) Potential of Rhizobium and Bradyrhizobium species as
plant growth promoting rhizobacteria on non-legumes: effect
on radishes (Raphanus sativus L.). Plant Soil 204:5767
Askeland RA, Morrison SM (1983) Cyanide production by
Pseudomonas fluorescens and Pseudomonas aeruginosa. Appl
Environ Microbiol 45:18021807
Axelsson R, Oxelfelt F, Lindblad P (1999) Transcriptional regula-
tion of Nostoc uptake hydrogenase. FEMS Microbiol Lett 170:
7781
Bakker AW, Schippers B (1987) Microbial cyanide production in
the rhizosphere in relation to potato yield reduction and Pseudo-
monas spp-mediated plant growth-stimulation. Soil Biol Bio-
chem 19:451457
Becker S, Holighaus G, Gabrielczyk T, Unden G (1996) O2 as the
regulatory signal for FNR-dependent gene regulation in Esche-
richia coli. J Bacteriol 178:45154521
Blumer C, Heeb S, Pessi G, Haas D (1999) Global GacA-steered
control of cyanide and exoprotease production in Pseudomonas
fluorescens involves specific ribosome binding sites. Proc Natl
Acad Sci USA 96:1407314078
Brysk MM, Lauinger C, Ressler C (1969) Biosynthesis of cyanide
from [2-14C-15N]glycine in Chromobacterium violaceum.
Biochim Biophys Acta 184:583588
Bunch AW, Knowles CJ (1982) Production of the secondary
metabolite cyanide by extracts of Chromobacterium vio-
laceum. J Gen Microbiol 128:26752680
Castric PA (1975) Hydrogen cyanide, a secondary metabolite of
Pseudomonas aeruginosa. Can J Microbiol 21:613618
Castric PA (1977) Glycine metabolism by Pseudomonas aerugi-
nosa: hydrogen cyanide biosynthesis. J Bacteriol 130:826831
Castric PA (1981) The metabolism of hydrogen cyanide by bacte-
ria. In: Vennesland B, Conn EE, Knowles CJ, Westley J, Wissing
F (eds) Cyanide in biology. Academic, London, pp 233261
Castric PA (1983) Hydrogen cyanide production by Pseudomonas
aeruginosa at reduced oxygen levels. Can J Microbiol 29:
13441349
Castric PA (1994) Influence of oxygen on the Pseudomonas aeru-
ginosa hydrogen cyanide synthase. Curr Microbiol 29:1921
Castric PA, Ebert RF, Castric KF (1979) The relationship between
growth phase and cyanogenesis in Pseudomonas auruginosa.
Curr Microbiol 2:287292
Castric PA, Castric KF, Meganathan R (1981) Factors influencing
the termination of cyanogenesis in Pseudomonas aeruginosa.
In: Vennesland B, Conn EE, Knowles CJ, Westley J, Wissing
F (eds) Cyanide in biology. Academic, London, pp 263274
Chlumsky LJ, Zhang L, Jorns MS (1995) Sequence analysis of sar-
cosine oxidase and nearby genes reveals homologies with key
enzymes of folate one-carbon metabolism. J Biol Chem 270:
1825218259
Cunningham L, Pitt M, Williams HD (1997) The cioAB genes
from Pseudomonas aeruginosa code for a novel cyanide-insen-
sitive terminal oxidase related to the cytochrome bd quinol ox-
idases. Mol Microbiol 24:579591
Evered D, Harnett S (1988) Cyanide compounds in biology. Ciba
Foundation Symposium 140. Wiley, Chichester
Flaishman MA, Eyal Z, Zilberstein A, Voisard C, Haas D (1996)
Suppression of Septoria tritici blotch and leaf rust of wheat by
recombinant cyanide-producing strains of Pseudomonas
putida. Mol Plant-Microbe Interact 9:642645
Gallagher L, Manoil C (1999) Pseudomonas aeruginosa kills
Caenorhabditis elegans by a multifactorial toxicity that re-
quires cyanide. In: Pseudomonas 99: biotechnology and
pathogenesis, abstract 75. American Society of Microbiology
Goldfarb WB, Margraf H (1967) Cyanide production by Pseudo-
monas aeruginosa. Ann Surg 165:104110
Hafner EW, Wellner D (1979) Reactivity of the imino acids
formed in the amino acid oxidase reaction. Biochemistry 18:
411417

Hildebrand U, Taraz K, Budzikiewicz H, Korth H, Pulverer G
(1985) Dicyano-bis-(pyridin-2,6-dicarbothioato)-ferrat(II)/fer-
rat(III), ein weiteres eisenhaltiges Redoxsystem aus der Kul-
turlsung eines Pseudomonas-Stammes. Z Naturforsch 40c:
201207
Hjberg O, Schnider U, Winteler HV, Srensen J, Haas D (1999)
Oxygen-sensing reporter strain of Pseudomonas fluorescens
for monitoring the distribution of low-oxygen habitats in soil.
Appl Environ Microbiol 65:40854093
Jones DA (1998) Why are so many food plants cyanogenic? Phy-
tochemistry 47:155162
Kiley PJ, Beinert H (1999) Oxygen sensing by the global regula-
tor, FNR: the role of the iron-sulfur cluster. FEMS Microbiol
Rev 22:341352
Knowles CJ (1988) Cyanide utilization and degradation by mi-
croorganisms. In: Evered D, Harnett S (eds) Cyanide com-
pounds in biology. Ciba Foundation Symposium 140. Wiley,
Chichester, pp 315
Knowles CJ, Bunch AW (1986) Microbial cyanide metabolism.
Adv Microb Physiol 27:73111
Kralik CA, Castric PA (1979) Respiration and cyanogenesis in
Pseudomonas aeruginosa. Curr Microbiol 3:7174
Kunz DA, Chen JL, Pan G (1998) Accumulation of -keto acids as
essential components in cyanide assimilation by Pseudomonas
fluorescens NCIMB 11764. Appl Environ Microbiol 64:
44524459
Laville J, Voisard C, Keel C, Maurhofer M, Dfago G, Haas D
(1992) Global control in Pseudomonas fluorescens mediating
antibiotic synthesis and suppression of black root rot of to-
bacco. Proc Natl Acad Sci USA 89:15621566
Laville J, Blumer C, Von Schroetter C, Gaia V, Dfago G, Keel C,
Haas D (1998) Characterization of the hcnABC gene cluster en-
coding hydrogen cyanide synthase and anaerobic regulation by
ANR in the strictly aerobic biocontrol agent Pseudomonas flu-
orescens CHA0. J Bacteriol 180:31873196
Leuthner B, Leutwein C, Schulz H, Horth P, Haehnel W, Schiltz
E, Schagger H, Heider J (1998) Biochemical and genetic char-
acterization of benzylsuccinate synthase from Thauera aromat-
ica: a new glycyl radical enzyme catalysing the first step in
anaerobic toluene metabolism. Mol Microbiol 28:615628
Michaels R, Hankes LV, Corpe WA (1965) Cyanide formation by
nonproliferating cells of Chromobacterium violaceum. Arch
Biochem Biophys 111:121125
Nazly N, Collins PA, Knowles CJ (1981) Cyanide production by
harvested Chromobacterium violaceum. In: Vennesland B,
Conn EE, Knowles CJ, Westley J, Wissing F (eds) Cyanide in
biology. Academic, London, pp 289299
Nicolet Y, Piras C, Legrand P, Hatchikian CE, Fontecilla-Camps
JC (1999) Desulfovibrio desulfuricans iron hydrogenase:the
structure shows unusual coordination to an active site Fe binu-
clear center. Structure 7:1323
Nishiya Y, Imanaka T (1998) Purification and characterization of
a novel glycine oxidase from Bacillus subtilis. FEBS Lett 438:
263266
Pierik AJ, Hulstein M, Hagen WR, Albracht SP (1998) A low-spin
iron with CN and CO as intrinsic ligands forms the core of the
active site in [Fe]-hydrogenases. Eur J Biochem 258:572578
Pierik AJ, Roseboom W, Happe RP, Bagley KA, Albracht SP
(1999) Carbon monoxide and cyanide as intrinsic ligands to
iron in the active site of [NiFe]-hydrogenases. J Biol Chem
274:33313337
Raybuck SA (1992) Microbes and microbial enzymes for cyanide
degradation. Biodegradation 3:318
Reimmann C, Beyeler M, Latifi A, Winteler H, Foglino M, Laz-
dunski A, Haas D (1997) The global activator GacA of
Pseudomonas aeruginosa PAO positively controls the produc-
tion of the autoinducer N-butyryl-homoserine lactone and the
formation of the virulence factors pyocyanin, cyanide, and li-
pase. Mol Microbiol 24:309319
Rhodius VA, Busby SJW (1998) Positive activation of gene ex-
pression. Curr Opin Microbiol 1:152159
Rodgers PB, Knowles CJ (1978) Cyanide production and degrada-
tion during growth of Chromobacterium violaceum. J Gen Mi-
crobiol 108:261267
177
Sawers RG (1991) Identification and molecular characterization of
a transcriptional regulator from Pseudomonas aeruginosa
PAO1 exhibiting structural and functional similarity to the
FNR protein of Escherichia coli. Mol Microbiol 5:14691481
Sawers G (1999) Biochemistry, physiology and molecular biology
of glycyl radical enzymes. FEMS Microbiol Rev 22:543551
Schmitz O, Boison G, Hilscher R, Hundeshagen B, Zimmer W,
Lottspeich F, Bothe H (1995) Molecular biological analysis of
a bidirectional hydrogenase from cyanobacteria. Eur J Bio-
chem 233:266276
Solomonson LP (1981) Cyanide as a metabolic inhibitor. In: Ven-
nesland B, Conn EE, Knowles CJ, Westley J, Wissing F (eds)
Cyanide in biology. Academic, London, pp 1128
Spiro S (1994) The FNR family of transcriptional regulators. An-
tonie van Leeuwenhoek 66:2336
Sun X, Ollagnier S, Schmidt PP, Atta M, Mulliez E, Lepape L,
Eliasson R, Graslund A, Fontecave M, Reichard P, Sjoberg
BM (1996) The free radical of the anaerobic ribonucleotide re-
ductase from Escherichia coli is at glycine 681. J Biol Chem
271:68276831
Unden G, Schirawski J (1997) The oxygen-responsive transcrip-
tional regulator FNR of Escherichia coli:the search for signals
and reactions. Mol Microbiol 25:205210
Vennesland B, Conn EE, Knowles CJ, Westley J, Wissing F
(1981a) Cyanide in biology. Academic, London
Vennesland B, Pistorius EK, Gewitz HS (1981b) HCN production
by microalgae. In: Vennesland B, Conn EE, Knowles CJ,
Westley J, Wissing F (eds) Cyanide in biology. Academic,
London, pp 349362
Vining LC (1990) Functions of secondary metabolites. Annu Rev
Microbiol 44:395427
Voisard C, Keel C, Haas D, Dfago G (1989) Cyanide production
by Pseudomonas fluorescens helps suppress black root rot of
tobacco under gnotobiotic conditions. EMBO J 8:351358
Voisard C, Bull CT, Keel C, Laville J, Maurhofer M, Schnider U,
Dfago G, Haas D (1994) Biocontrol of root diseases by
Pseudomonas fluorescens CHA0:current concepts and experi-
mental approaches. In: OGara F, Dowling DN, Boesten B
(eds) Molecular ecology of rhizosphere microorganisms. VCH,
Weinheim, pp 6789
Von Tigerstrom M, Campbell JJR (1966) The accumulation of -
ketoglutarate by suspensions of Pseudomonas aeruginosa. Can
J Microbiol 12:10051013
Wierenga RK, Terpstra P, Hol WG (1986) Prediction of the occur-
rence of the ADP-binding -fold in proteins, using an amino
acid sequence fingerprint. J Mol Biol 187:101107
Wilson MT, Antonini G, Malatesta F, Sarti P, Brunori M (1994)
Probing the oxygen binding site of cytochrome c oxidase by
cyanide. J Biol Chem 269:2411424119
Winson MK, Camara M, Latifi A, Foglino M, Chhabra SR,
Daykin M, Bally M, Chapon V, Salmond GPC, Bycroft BW,
Lazdunski A, Stewart GSAB, Williams P (1995) Multiple N-
acyl-L-homoserine lactone signal molecules regulate produc-
tion of virulence determinants and secondary metabolites in
Pseudomonas aeruginosa. Proc Natl Acad Sci USA 92:9427
9431
Wissing F (1974) Cyanide formation from oxidation of glycine by
a Pseudomonas species. J Bacteriol 117:12891294
Wissing F (1975) Cyanide production from glycine by a homo-
genate from a Pseudomonas species. J Bacteriol 121:695699
Wissing F, Andersen KA (1981) The enzymology of cyanide pro-
duction from glycine by a Pseudomonas species. In: Vennes-
land B, Conn EE, Knowles CJ, Westley J, Wissing F (eds)
Cyanide in biology, Academic, London, pp 275288
Zanker H, Lurz G, Langridge U, Langridge P, Kreusch D,
Schrder J (1994) Octopine and nopaline oxidases from Ti
plasmids of Agrobacterium tumefaciens: molecular analysis,
relationship, and functional characterization. J Bacteriol 176:
45114517
Zimmermann A, Reimmann C, Galimand M, Haas D (1991)
Anaerobic growth and cyanide synthesis of Pseudomonas
aeruginosa depend on anr, a regulatory gene homologous with
fnr of Escherichia coli. Mol Microbiol 5:14831490