-
Vol. 173, No. 7JOURNAL OF BACTERIOLOGY, Apr. 1991, p.
2311-23180021-9193/91/072311-08$02.00/0Copyright © 1991, American
Society for Microbiology
Phosphorylation of the AfsR Product, a Global RegulatoryProtein
for Secondary-Metabolite Formation in
Streptomyces coelicolor A3(2)SOON-KWANG HONG, MORIKAZU KITO,t
TERUHIKO BEPPU, AND SUEHARU HORINOUCHI*
Department ofAgricultural Chemistry, Faculty of Agriculture, The
University of Tokyo,Bunkyo-ku, Tokyo 113, Japan
Received 29 October 1990/Accepted 31 January 1991
The AfsR protein is essential for the biosynthesis at the
wild-type level of A-factor, actinorhodin, andundecylprodigiosin in
Streptomyces coelicolor A3(2) and Streptomyces lividans. Because
overexpression of theafsR gene caused some deleterious effect on
these strains, a multicopy plasmid carrying the whole afsR gene
wasintroduced into Streptomyces griseus, from which a crude cell
lysate was prepared as a protein source. The AfsRprotein was
purified to homogeneity from the cytoplasmic fraction through
several steps of chromatography,including affinity column
chromatography with ATP-agarose and use of anti-AfsR antibody for
its detection.The molecular weight of AfsR was estimated by sodium
dodecyl sulfate-polyacrylamide gel electrophoresis andby gel
filtration to be 105,300, which is in good agreement with that
deduced from the nucleotide sequence ofafsR. The purified AfsR
protein was found to be phosphorylated through the transfer of the
y-phosphate groupof ATP in the presence of the cell extracts of S.
coelicolor A3(2) and S. lividans. This phosphorylation
proceededvery rapidly, and no competition was observed with CTP,
GTP, UTP, or cyclic AMP. In the cell extract of S.griseus, no
activity phosphorylating the AfsR protein was detected, suggesting
that this activity is not generallypresent in Streptomyces spp. but
is specific to certain species. It is conceivable that the extent
of phosphorylationof the AfsR protein modulates its regulatory
activity which, in turn, regulates expression of some target
gene(s)involved in the secondary-metabolite formation in S.
coelicolor A3(2).
The afsR gene is a pleiotropic and essential regulatorygene for
secondary metabolism in Streptomyces coelicolorA3(2) and a related
species, Streptomyces lividans (11, 17,33). Introduction of the
gene cloned on a plasmid into S.lividans caused marked production
of the pigmented antibi-otics actinorhodin and undecylprodigiosin
(16). The induc-tion of actinorhodin production was found to occur
throughtranscriptional stimulation of the genes involved in
theantibiotic biosynthesis (20). Nucleotide sequencing of theafsR
gene revealed that it codes for a 993-amino-acid proteinwith a
molecular weight of 105,600 (18). The AfsR proteincontains A- and
B-type ATP-binding consensus sequences atits NH2-terminal portion
and two DNA-binding consensussequences with a helix-turn-helix
motif at its COOH-termi-nal portion. A mutation of either of the
two ATP-bindingconsensus sequences, which was generated by
site-directedmutagenesis, resulted in the loss of the ability of
afsR tocause production of pigments in S. lividans. In
addition,disruption of the chromosomal afsR gene by use of
thecloned afsR gene and phage 4C31KC515 resulted in asignificant
loss of pigment production in S. coelicolor A3(2),which indicates
that afsR has an obligatory role in normalantibiotic synthesis
(18).
Historically, the afsR gene was identified because it
com-plemented an afsB mutation on the chromosome (16, 17).Although
the cloned afsR gene was assumed to coincide withthe afsB gene,
subsequent experiments have shown that theafsR product is a bypass
function with regard to afsBcomplementation (33). More recently,
the AfsR protein hasbeen shown to be important in its own right,
since disruption
* Corresponding author.t Present address: Applied Research
Laboratories, Ajinomoto
Co., Kawasaki-ku, Kanagawa 210, Japan.
of its function results in a loss of pigment production
(18).Each of the NH2- and COOH-terminal halves could
partiallyconfer pigment production on S. lividans.The above
observations prompted us to examine the
possible ability of the AfsR product to interact with ATP,which
might be associated with its regulatory function. Forthis purpose,
we purified the AfsR protein to homogeneityand determined its
characteristics. In this paper, we describethe purification of the
AfsR protein and the phosphorylationof it by a phosphokinase
activity present in S. coelicolorA3(2) and S. lividans.
MATERIALS AND METHODS
Materials. Restriction endonucleases, DNA-modifying en-zymes,
were purchased from Takara Shuzo Co., Ltd., andwere used according
to the recommendation of the supplier.[_y-32P]ATP (5,000 Ci/mmol),
[a-32P]ATP (3,000 Ci/mmol),and an oligonucleotide-directed
mutagenesis kit were ob-tained from Amersham Corp. ATP, CTP, GTP,
UTP, cyclicAMP, and ATP-agarose (AGATP type 4) were purchasedfrom
Pharmacia, Inc., and cellulose-nitrate membranes werefrom
Schleicher & Schuell Corp. The nonionic detergentThesit, which
has the structural formula dodecylpoly(ethyl-ene glycol ether)n,
was supplied by Boehringer GmbH.Thiostrepton was a gift from Asahi
Chemical Industry,Shizuoka, Japan.
Strains and plasmids. The bacterial strains and plasmidsused are
listed in Table 1.Recombinant DNA work. Preparation of plasmid
DNA,
recombinant DNA work, and protoplast transformationwere
performed as previously described (14). For expressionof the AfsR
protein at a higher level, plasmid pIJ702-C81 wasconstructed (see
Fig. 1). Plasmid pIJ702 (25), a high-copy-
2311
on April 7, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
2312 HONG ET AL.
TABLE 1. Bacterial strains and plasmids
Designation Relevant characteristics Reference
Streptomyces strainsS. lividans TK21 15S. coelicolor A3(2) hisAl
uraAl strAl Scp-1- 5M130 Scp-2-; A-factor producer
S. griseus HH1 A-factor-deficient mutant 19derived from S.
griseusIFO 13350
Escherichia coli F- hsdS20 recA131 ara-14HB101 proA2 lacYl galK2
rpsL20
supE44
PlasmidspIJ41 Neomycin and thiostrepton 4
resistance; copy number,3-4/chromosome
pIJ702 Thiostrepton resistance; 25melanin'; copy
number,40-300/chromosome
pIJ41-AP3 Thiostrepton resistance; 17contains the afsR gene
pIJ702-C81 Thiostrepton resistance; This studycontains the afsR
gene
pCA1 Ampicillin resistance; 18contains a prochymosin-afsR fused
gene
number plasmid (40 to 300 copies per chromosome) inStreptomyces
spp., was digested with PstI and BglII restric-tion enzymes, and
the larger fragment (5.3 kb) was purifiedby agarose gel
electrophoresis. For preparation of a DNAfragment containing the
promoter and whole coding regionof the afsR gene by cleavage with
restriction enzymes ofpIJ41-AP3, a SmaI site at nucleotide position
1 reported inreference 18 was changed into a PstI site by attaching
ahexamer PstI linker. Then, a 3.8-kb PstI-BclI fragmentcontaining
the whole afsR gene was ligated with PstI-plus-BglII-digested
pIJ702. The ligation mixture was introducedby transformation into
protoplasts of Streptomyces griseusHH1. Transformants were selected
on medium containing 40,g of thiostrepton per ml.Media. The liquid
medium for S. griseus HH1 had a pH of
7.2 and contained, in grams per liter, yeast extract
(DifcoLaboratories), 2; meat extract (Kyokuto), 2;
Bacto-Peptone(Difco), 4; NaCl, 5; MgSO4 7H20, 2; and glucose,
10.R2YE medium (14) containing 0.2% asparagine instead ofproline
was used for regeneration of protoplasts. For theliquid medium of
S. coelicolor A3(2) M130, R2YE mediumwithout agar was used. YMPG
medium (pH 7.2) (containing,in grams per liter, yeast extract, 4;
malt extract [Difco], 10;MgCl2 .6H20, 2; Bacto-Peptone, 1; and
glucose, 10) wasused for the cultivation of S. lividans TK21.
Preparation of anti-AfsR protein antibody. E. coli
HB101harboring plasmid pCA1 (18) was used to produce
theprochymosin-AfsR fused protein (1,323 amino acids; Ala-1to
Leu-361 of prochymosin and Ala-43 to Arg-993 of AfsR).The strain
was grown in M9 medium supplemented with 2%Casamino Acids and 0.5%
glucose to an optical density at600 nm of about 0.5, and
3-,-indolylacrylic acid was addedto a final concentration of 15
pug/ml. After incubation for 15h with shaking at 37°C, cells were
harvested, washed with 50mM Tris - Cl (pH 7.5) buffer, and
disrupted by sonication.The fused protein overproduced in the form
of inclusion
bodies was collected by low-speed centrifugation at 5,000 xg for
10 min. The precipitate was suspended in the samebuffer, and the
proteins were applied onto sodium dodecylsulfate
(SDS)-polyacrylamide (7% acrylamide) gels for elec-trophoresis. The
145-kDa protein corresponding to the pro-chymosin-AfsR fused
protein was electroeluted from the gel(22), and 500 pLg of the
protein was injected into two rabbits(New Zealand White rabbits,
male; body weight, 2 kg). Therabbits were booster injected 1 week
later with a further 500ptg of the protein and were bled 4 weeks
after the lastinjection. The immunoglobulin G fraction was purified
fromthe antiserum by protein A-affinity column chromatography(26).
The antibodies specific to prochymosin in the immuno-globulin G
fraction were removed by immunoprecipitation(3) with prochymosin,
and the supernatant which did notshow any positive reaction against
prochymosin was used asthe anti-AfsR antibody.
Purification of the AfsR protein. Throughout the purifica-tion
procedure, the AfsR protein was identified by SDS-polyacrylamide
gel electrophoresis and Western immunoblotanalysis (6) with the
anti-AfsR protein antibody. All of theoperations were performed at
4°C, except fast protein liquidchromatography (FPLC) operations
were performed at roomtemperature. The amount of protein was
measured by themethod of Lowry et al. (29).
(i) Preparation of crude extract. A 3-day-old culture of
S.griseus HH1 containing plasmid pIJ702-C81 in the presenceof 20
,ug of thiostrepton per ml was diluted 1:10 into freshliquid medium
and grown at 30°C for 3 days on a reciprocalshaker. Mycelia (wet
weight, 210 g) obtained from 10 liters ofculture were washed with
buffer A (10 mM Tris Cl [pH6.3], 1 mM EDTA, 1 mM
,B-mercaptoethanol) and suspendedin 1.5 liters of buffer A. They
were disrupted by sonication.Cell debris was removed by
centrifugation for 1 h at 20,000x g in a high-speed centrifuge.
Proteins in the supernatantfluid that were precipitated with 33%-
to 55%-saturatedammonium sulfate were collected by centrifugation
at 20,000x g for 1 h. The precipitate was dissolved in 500 ml of
bufferA and dialyzed against 20 liters of the same buffer
overnightwith three changes of the buffer.
(ii) DEAE-Toyo Pearl column chromatography. After re-moval of
precipitates of the dialyzed sample by centrifuga-tion, the
supernatant was diluted with buffer A to a finalvolume of 2 liters.
This sample solution was applied to aDEAE-Toyo Pearl column (Toyo
Soda Manufacturing Co.,Ltd., Tokyo, Japan; 4 by 40 cm) previously
equilibrated withbuffer A. The column was washed with the same
buffer andwashed again with buffer A containing 0.1 M NaCl until
nomore protein was eluted. Buffer A containing 0.2 M NaClwas then
applied to elute the AfsR protein. The fractionscontaining the AfsR
protein were pooled and dialyzedagainst 5 liters of buffer B (10 mM
Tris Cl [pH 7.0], 5 mMEDTA, 5 mM ,B-mercaptoethanol) overnight with
twochanges of the buffer.
(iii) Affinity chromatography with an ATP-agarose column.After
centrifugation to remove a small amount of the precip-itates in the
dialysate, the supernatant was applied to anATP-agarose column (1.5
by 10 cm) (28) previously equili-brated with buffer B. After the
column had been washed withthe same buffer, proteins were eluted
with a linear NaClgradient from 0 to 1 M in buffer B containing
0.01% Thesit,a nonionic detergent, to prevent the aggregation
betweenprotein molecules. The fractions from 0.3 to 0.45 M
NaClgradient portions containing the AfsR protein were pooled.
(iv) Phenyl-superose column chromatography. The samplethus
obtained was brought to 20% saturation of ammonium
J. BACTERIOL.
on April 7, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
PHOSPHORYLATION OF THE AfsR PROTEIN 2313
sulfate. After filtration through a 0.22-,um-pore-size mem-brane
filter, it was applied to a phenyl-superose column (0.5by 5 cm)
previously equilibrated with buffer B containing20% ammonium
sulfate. The AfsR protein was eluted with alinear ammonium sulfate
gradient from 20 to 0% in the samebuffer containing 0.01% Thesit in
a Pharmacia FPLC system.
(v) Gel permeation chromatography. For further purifica-tion,
superose-6 and superose-12 were connected in theFPLC system and gel
permeation chromatography of theAfsR protein was performed twice in
succession, once inbuffer C (10 mM Tris Cl [pH 7.4], 1 mM EDTA, 1
mMP-mercaptoethanol) containing 0.01% Thesit and once inbuffer C
only to remove the nonionic detergent. The AfsRprotein purified in
this way was stored in 50% glycerol at-800C.Preparation of crude
extracts as a source of phosphokinases.
S. coelicolor A3(2) M130, S. lividans TK21 and S. griseusHH1
were cultivated at 30°C for 3 days on a reciprocalshaker, and the
mycelia were harvested. After the myceliahad been washed with
buffer C, they were suspended in thesame buffer and disrupted by
sonication. Cell debris wasremoved by centrifugation at 35,000 x g
for 30 min, and thesupernatants were used as a mediator for
phosphorylation ofthe AfsR protein.
Phosphorylation protocol. The purified AfsR protein
wasphosphorylated with the crude extracts of S. coeli-olorA3(2)
M130, S. lividans TK21, or S. griseus HH1 as aphosphokinase
preparation in a mixture containing 10 mMTris * Cl (pH 7.4), 4 mM
MgCl2 - 6H20, 3 mM P-mercapto-ethanol, 2 mM MnCl2, 1 mM EDTA, and
10 ,uCi of[_y-32P]ATP, mainly according to the conditions described
byCherry et al. (8). After 5 min of incubation at room
temper-ature, the reaction was terminated by the addition,
withboiling for 2 min, of the SDS sample buffer for
electropho-resis to give a 0.1% concentration of SDS.
Phosphorylatedproteins were separated by 0.1% SDS-10%
polyacrylamidegel electrophoresis and detected by autoradiography
at-80°C with a Du Pont Cronex intensifying screen.Polyacrylamide
gel electrophoresis. An SDS-polyacryl-
amide gel (10% acrylamide) system (27) was used to examinethe
purity of the AfsR preparation and to determine itsmolecular
weight. A polyacrylamide gel system (10% acryl-amide) without SDS
was also used (10).
RESULTS
Purification of the AfsR protein. For overexpression of theAfsR
protein, we constructed pIJ702-C81 containing thewhole afsR gene on
the multicopy plasmid vector pIJ702(Fig. 1). As previously reported
(18), neither S. coelicolorA3(2) nor S. lividans allowed the
replication of a multicopyplasmid carrying the afsR gene, probably
because the pres-ence of afsR at a high copy number causes
production ofactinorhodin and undecylprodigiosin to such an extent
thatthese pigmented antibiotics led to the death of the host
cells.The afsR gene, even on pIJ41, whose copy number is 3 to 4per
chromosome (14), caused a decrease in the growth rateof S.
lividans. We therefore introduced pIJ702-C81 into S.griseus HH1,
which does not contain the biosynthetic genesfor these antibiotics;
from this a crude extract as the startingmaterial for purification
of the AfsR protein was prepared. Inaddition, S. griseus HH1 was
expected to produce noprotein homologous to the AfsR protein,
because it containsno DNA sequence homologous to the afsR gene
(21). Al-though the growth of S. griseus HH1 containing
pIJ702-C81
Ai
plJ702-C81 #
PstlI afsR cii/BgIII
I IfsR 1T D A-idn
ATP-binding DNA-bindingsequence sequence
FIG. 1. Construction of plasmid pIJ702-C81 carrying the
afsRgene. The single line of the circle represents the sequence of
plasmidpIJ702 containing the thiostrepton resistance gene (tsr) and
a mela-nin production gene (mel), and the open bar represents the
DNAfragment containing the S. coelicolor A3(2) afsR gene. Just
up-stream of the coding sequence, there are two transcriptional
startpoints (18), as shown by an arrow. The AfsR protein of 993
aminoacids contains two ATP-binding consensus sequences at its
NH2-terminal portion and two DNA-binding consensus sequences at
itsCOOH-terminal portion.
was considerably reduced, the addition of thiostrepton to
themedium allowed the plasmid to be maintained.Western blot
hybridization with the anti-AfsR protein
antibody was employed for detection of the AfsR protein.
Aprotein band giving positive hybridization in the Westernblot was
detected in the supernatant of the sonic extract ofmycelium, i.e.,
in the soluble fraction but not the membranefraction (Fig. 2C, lane
2). Starting with this crude lysate, wepurified the AfsR protein to
homogeneity by ammoniumsulfate fractionation and several steps of
column chromatog-raphy. Among these steps, the affinity column
chromatogra-phy with ATP-agarose was very useful for its
purification.The final preparation gave a single band on an
SDS-poly-acrylamide gel even by silver-staining (Fig. 2B). The
appar-ent molecular weight of the AfsR protein was estimated to
be105,300 from its mobility on SDS-polyacrylamide gel
elec-trophoresis. This value was in good agreement with that
(M,105,600) deduced from the nucleotide sequence of afsR (18).The
molecular weight of the native AfsR protein estimatedon the basis
of both SDS-polyacrylamide gel electrophoresiswith a nondenaturing
buffer (0.06 M Tris * Cl, pH 6.8, 10%glycerol) without a boiling
treatment (9) and superose-6 gelfiltration was exactly the same as
that determined on thedenaturing SDS-polyacrylamide gel. These data
show that
VOL. 173, 1991
on April 7, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
2314 HONG ET AL.
M 1 2 34 5 6 7 M 7 M 1 2 3
94 --IS"'_:.,.wr
67
43 w30
20.1 N- _
Ih
A BFIG. 2. Purification of the AfsR protein. (A) SDS
gel used for detection of the AfsR protein and detepurity; (B)
silver staining of the purified AfsR protein; (hybridization
performed with the anti-AfsR antibodytein is indicated by the
arrow. Lanes 1, Total-cell lysat4HH1 harboring pI702 as a control
(10 ,ug of prsupernatant of sonic extract of total-cell lysate
prgriseus HH1 harboring pLJ702-C81 (10 ,ug of proteinammonium
sulfate fractionation (5 ,ug of protein);DEAE-cellulose
chromatography (2.5 ,ug of proteinATP-agarose chromatography (1.5
,ug of protein); lane,superose chromatography (1 ,ug of protein);
lanes 7, asuperose-12 chromatography (1 ,ug of protein). The fclar
mass markers were used in lanes M: phosphorylbovine serum albumin
(67 kDa), ovalbumin (43 kDa),drase (30 kDa), and soybean trypsin
inhibitor (20.1 kI
4 5 6 7 the AfsR protein thus purified exists as a monomer in
thecytoplasmic fraction.
Phosphorylation of the AfsR protein. The amino acidsequence of
AfsR protein deduced from its nucleotide se-
- - quence indicated that it contains A- and B-type
ATP-bindingconsensus sequences at its NH2-terminal portion (18).
Byanalogy with two-component regulatory systems (32), wehad
expected that the AfsR protein might be autophospho-rylated and
that it could regulate secondary metabolismdepending on the extent
of phosphorylation or dephospho-rylation of AfsR. The purified AfsR
protein was thereforeassayed for autophosphorylation by incubating
it with[_y-32P]ATP for 5 min in the presence of 4 mM MgCl2.Contrary
to our expectation, however, no phosphorylationoccurred in this
assay system (Fig. 3, lane 1).We next examined phosphorylation of
AfsR protein with
-' the cell lysate of S. coelicolor A3(2) M130, from which
the0-polyacrylamide afsR gene was derived. Reaction conditions were
basedrmrnation of its mainly on those of Cherry et al. (8). A
protein with an(C) Western blot apparent molecular weight of
105,300, in addition to several1.The AfsR pro-efrom S. griseus
other proteins, was 32P labeled when the purified AfsRotein); lanes
2, protein was incubated with the cell lysate in the presence
of*epared from S. [y-32P]ATP (Fig. 3). When [ot-32P]ATP instead of
[_y-32P]ATP); lanes 3, after was used, no labeling occurred, as
described below. Thelanes 4, after degree of 32p labeling of this
protein was apparently propor-
); lanes 5, after tional to the amount of the AfsR protein
added. In addition,s 6, after phenyl- when the AfsR protein was not
added to the assay mixture,ifter superose-6- no detectable labeling
could be observed under these condi-lase b (94kDa)m tions (lane 6).
However, a very long exposure of the gel, carbonic anhy- showed a
faint signal in lane 6 at the position correspondingDa). to the
AfsR protein (data not shown). All of these data
.---:?, -- Z~- CC 0 0~- :_ - 2: 007 -L;me
'lurilf'ied A tf;i iAnnle ) ' .2 -. -*(i. I(ico1(3r
3 vext rac t. ( lg 1
Cr- F))AiT ( uI )
zf- _. 2DI
_ :-xZ~:- :- .-- OI
% -'.3 % I %.
94
67
4330
20.1
A BFIG. 3. Phosphorylation of the AfsR protein mediated by the
crude extract from S. coelicolor A3(2) M130. Reaction mixtures
containing
[-y-32P]ATP and various amounts of the purified AfsR protein and
the cell extract were incubated for 5 min at room temperature. The
mixtureswere then separated by SDS-polyacrylamide gel
electrophoresis, and the autoradiogram (A) and Western blot (B)
were obtained. The two32P-labeled bands indicated by arrowheads are
apparently degraded products of the 32P-labeled AfsR protein. The
phosphorylated AfsRprotein is indicated by the arrow. Molecular
masses (in kilodaltons) are indicated on the left.
J. BACTERIOL.
*4f-
A:
on April 7, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
PHOSPHORYLATION OF THE AfsR PROTEIN 2315
clearly show that the AfsR protein is phosphorylated when itis
incubated with ATP and the cell extract of S. coelicolorA3(2)
cells. The experiments with a lesser amount of the celllysate, as
described below, showed that this phosphorylationwas very rapid and
that the dephosphorylation was alsorapid; this is an explanation
for the lack of a comparableincrease of 32p signals of
corresponding proteins with eitheradded AfsR or various amounts of
cell extract.
Concerning the additional 32P-labeled proteins, two dis-tinct
bands, indicated by arrowheads in Fig. 3, are appar-ently degraded
products of the AfsR protein, since thereaction mixture without
exogenous AfsR (lane 6) does notyield these bands. This is also
supported by the Western blothybridization which shows that these
are immunoreactivewith the anti-AfsR antibody. On the other hand,
other32P-labeled bands observed are assumed to be proteins thatare
32p labeled by this reaction system. Because no 32plabeling of
these proteins occurs with [a-32P]ATP, as de-scribed below, this
implies that the incubation of the cellextract with ATP under these
conditions yields phosphory-lated proteins in addition to the AfsR
protein, indicating thepresence of a protein phosphokinase(s). It
is not clear atpresent whether the phosphokinase activity
responsible forthe phosphorylation of AfsR also phosphorylates some
ofthese proteins.The above assay also suggested that the content of
the
AfsR protein in S. coelicolor A3(2) cells was too low to
bedetected by this assay system, although the AfsR proteinthat had
already been phosphorylated could not be detectedby this assay.
Consistent with this, an immunoblot assaywith the anti-AfsR
antibody failed to show any proteinsreactive with the antibody,
even in concentrated cell lysatesprepared from 3- or 4-day-old
culture, probably because ofthe small number of the AfsR
molecules.
Effects of various nucleotides on phosphorylation of theAfsR
protein. As mentioned above, incubation of the AfsRprotein with the
cell lysate of S. coelicolor A3(2) in thepresence of [ct-32P]ATP
instead of [_y-32P]ATP did not yield a32P-labeled AfsR protein
(Fig. 4, lane 3). The 32p labeling ofthe AfsR protein with
[_-32P]ATP was competitively inhib-ited by an excess of unlabeled
ATP (Fig. 4, lane 4). Theseresults show that the y-phosphate group
of ATP is trans-ferred to the AfsR protein. We next examined the
effects ofvarious nucleotides in the reaction mixture on
phosphoryla-tion. Unlike ATP, an excess of CTP, GTP, UTP, or
cyclicAMP caused no significant effect on the phosphorylation ofthe
AfsR protein with [y-32P]ATP (Fig. 4, lanes 5 to 8).
Kinetics of phosphorylation of the AfsR protein. The timecourse
of phosphorylation of the AfsR protein in the abovesystem was
examined. The reaction occurred very rapidly,and most of the AfsR
molecules in the reaction mixtureappeared to be phosphorylated
within 20 s (Fig. 5A). Theamount of phosphorylated AfsR protein
continued to in-crease for 2.5 min. After 5 min, it gradually
decreased,presumably because the reaction reached equilibrium
andthen phosphatase and protease in the cell lysate degraded
thephosphorylated AfsR protein.Because of the unexpectedly rapid
phosphorylation, we
used a lesser amount of the cell lysate in the reaction
mixturein order to confirm that the reaction was linear with
time(Fig. SB). The amount of 32P-labeled AfsR gradually in-creased
with the incubation period, indicating clearly thatthe
phosphorylation is linear with time.
Presence of a kinase activity specific to AfsR in S. lividansbut
not in S. griseus. A phosphokinase activity similar to thatin S.
coelicolor A3(2) was expected to exist in S. lividans,
1 234 5 6 7 8
94
67
43
30
20.19
FIG. 4. Autoradiogram showing the effects of various
nucleo-tides on phosphorylation of the AfsR protein. The standard
reactionmixture (lane 2) in a final volume of 20 ,ul containing 10
,uCi of[y-32P]ATP, 2.2 pmol of the purified AfsR protein, and the
cellextract (10 ,ug of protein) of S. coelicolor A3(2) was
incubated for 5min at room temperature. The cell extract was
omitted from thestandard mixture as a control (lane 1). [_y-32P]ATP
was replaced by[a-32P]ATP (lane 3). Lanes 4 to 8 show the
phosphorylation con-taining 20 nmol (each) of ATP (lane 4), CTP
(lane 5), GTP (lane 6),UTP (lane 7), and cyclic AMP (lane 8).
Molecular masses (inkilodaltons) are indicated on the left.
since introduction of the afsR gene in this strain
causedproduction of actinorhodin and undecylprodigiosin in
verylarge amounts. Examination for a phosphokinase capable
ofphosphorylating the AfsR protein clearly showed the pres-ence of
such an activity in the cell lysate of S. lividans TK21(Fig. 6,
lanes 2 to 9). When the reaction mixture (lane 7)containing the
cell lysate of S. lividans and [_y-32P]ATP withno exogenous supply
of the purified AfsR protein wasanalyzed similarly by means of a
prolonged exposure, a faint
1 2 3 4 5 6
w
94 _..
67
4330 _20.1 .-
A
1 2 3 4 5 6
BFIG. 5. Autoradiogram showing phosphorylation of the AfsR
protein depending on reaction time. Each lane contained 2.2 pmol
ofthe AfsR protein in 20 ,ul of the reaction mixture. The
reactionmixture was incubated for 0 s (lanes 1), 20 s (lanes 2), 1
min (lanes3), 2.5 min (lanes 4), 5 min (lanes 5), and 20 min (lanes
6) at roomtemperature and immediately subjected to gel
electrophoresis. (A)Crude extract (10 ,ug of protein) from S.
coelicolor A3(2) M130 and10 ,uCi of [-y-32P]ATP; (B) crude extract
(4 ,g of protein) and 20 ,uCiof [y-32P]ATP. Molecular masses (in
kilodaltons) are indicated onthe left.
VOL. 173, 1991
A0.
6:
W, 'M::im. :., '410..... .M"::-- :..
4'
on April 7, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
2316 HONG ETAL.J.BCEOL
-. N enCw u N.O ,- C OrO - N rl"tA UNr O"D - -4 N m A UgN O :l-
o OyO - N r.iJrc %nO rl-Lane ----- - - - .."
Puriried AfVsR (pmnole) N~ N NN N~N~N~N~ N~ -~ N" NN N N N N '4
N N N N Ny Nm - NmNNN N N N 0 NC N N N N NO0 - NN ' O
S. Iividanscrude extract (jug)
UN 0 LIN 0 UIN 0 0 0-- - N N N Lr U% UCN
UN% 0 UCN 0 UN% 0 0 0'-''-' NfyUNUN UN l
S. qrisouscrude extract (nug)
Cr-32PWrAp (LI)i
67-S
43"4W
UIN 0 UIN 0 UIN 0 0 0-. - Nl N% UN% UN% UN
1.''v-n am.
A B
FIG. 6. Phosphorylation of the AfsR protein by crude extracts
from S. lividans TK21 and S. griseus HH1: SDS-polyacrylamide gel of
the
phosphorylating reaction mixtures (A) and its autoradiogram (B).
Lanes 2 to 9 contain the crude extract of S. lividans HH21, and
lanes 10
to 17 contain the crude extract of S. griseus HH1. The reaction
was continued fo'r min at room temperature. The AfsR protein band
isindicated by the arrow. Molecular mas'ses (in kilodaltons) are
indicated on the left.
32P-labeled signal at the position corresponding to AfsR was
observed. This implies that in S. lividans the AfsR protein
which can be phosphorylated is produced in a very small
amount, as in the case of S. coelicolor A3(2). The other
proteins of S. lividans that were 32P labeled in this system
appeared to be the same as those in S. coelicolor A3(2). A
distinct band of a higher molecular weight than AfsR that is
seen in all of the lanes appears to be missing in lane 7 in
this
photograph, but a weak 32P signal corresponding to this
position is seen on the X-ray ffilm. These data reflect
theobservations that these two strains are very similar to each
other in many aspects.
On the other hand, a similar experiment with the cell
lysate of S. griseus HH1 showed the absence of a phosphoki-
nase activity specific to the AfsR protein (Fig. 6, lanes 10
to
17). The number and sizes of S. griseus proteins that were
32P labeled are different from those of S. coelicolor A3(2)
and
S. lividans proteins.
DISCUSSION
Despite the ability of both COOH- and NH2-terminal
portions to complement the afsB mutation, the large size of
the AfsR protein predicted by the nucleotide sequence has
been confirmed by this study. The present study also clearly
demonstrates that the AfsR protein accepts the y~-phosphateof
ATP when it is incubated with the cell extracts of S.
coelicolor A3(2) and S. lividans. It was suggested that the
phosphorylated A.fsR protein exerts its positive function,
probably by enhancing the transcription of its target
gene(s)
(20). The phosphokinase activity specific for AfsR appears
not to be present in general in Streptomyces spp. These
features of the AfsR protein, together with the presence of
a.
specific phosphorylating activity, remind us of so-called
two-component regulatory systems such as NtrB-NtrC (31,
34), EnvZ-OmpR (2, 24), PhoB-PhoR (30) and CheA-CheY
(12, 13). The two-component modulator-effector pairs are
very likely prevalent in a variety of procaryotes. The
genet-
ical and biochemical features so far observed for the AfsR
protein seem to conform with those of a transcriptional
regulator, called an effector, of the two-component systems,
although no significant similarity in amino acid sequence
between AfsR and other components of these systems (23) is
found. If so, the protein kinase activity for AfsR which we
have detected in this study is then specified by a protein,
a
modulator, which is autophosphorylated and is able to
transfer a phosphate group from itself to the AfsR protein.
Further extensive studies on the AfsR protein. and the
putative protein kinase are apparently required to reveal
the
mechanisms of regulation exerted by the pair of AfsR and
the A&fR-phosphorylating activity. We are now trying
topurify the kinase by following the AfsR-phosphorylating
activity.S. lividans growing without actinorhodin production
pro-
duces a very small amount of the AfsR counterpart that can
be phosphorylated (data not shown), just like the case of S.
coelicolor A3(2). Why does S. lividans produce actinorhodin
only under unusual cultural conditions, for example, on
media containing sucrose at a high concentration? Thedifference
in the profile of actinorhodin production between
these two strains is presumably ascribed to some other
downstream machinery to which the positive regulatory
LA 0 UN1 0 UN 0 0 0-. v- N N UN% UN UNA
J. BACTERIOL.
- - - 1- .. -4 - - - " - - - -4 -4 -- " -4 -4 -4 " q .4 .4 -4 -4
1.4 9-4 9-4 " qllq
on April 7, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
PHOSPHORYLATION OF THE AfsR PROTEIN 2317
signal of the phosphorylated AfsR is transferred. We spec-ulate
as a model that in the case of actinorhodin, thephosphorylated AfsR
protein binds the regulatory region ofactll controlling the
expression of the biosynthetic clusterand stimulates its
transcription. According to this specula-tion, the actll regulatory
region may possess some differ-ence from that of S. coelicolor
A3(2); as a result, a largeramount of AfsR is required for full
expression of actll thanin the case of S. coelicolor A3(2).
However, this speculationstill cannot answer the above
question.
In addition to afsR, afsB (20) and absA (1) also globallycontrol
the secondary metabolite formation in S. coelicolorA3(2). Both afsB
and absA mutations are circumventedwhen afsR is introduced on a
plasmid. Interestingly, amulticopy plasmid, pIJ702-AP22, containing
the DNA se-quence coding for only a COOH-terminal region of the
AfsRprotein is capable of complementing, to some extent,
thedeficiency in actinorhodin production in afsB and absAmutants
(7). We also showed that only an NH2-terminalregion of AfsR still
possessed the ability to confer actinorho-din production on S.
lividans (18). The experiments with themutated AfsR protein suggest
that the truncated AfsR pro-teins are not phosphorylated (data not
shown). We cannotfind any explanation for these interesting
phenomena, but itis clear that all three of these genes so far
identified arefunctioning for the secondary metabolism in S.
coelicolorA3(2) and that at least in the case of AfsR, the extent
ofphosphorylation is profoundly associated with the regula-tion.The
AfsR gene can be maintained at a high copy number in
S. griseus, although it affects the growth rate significantly.No
enhancement of streptomycin production by the pres-ence of afsR was
observed. It is clear that the AfsR proteinis not phosphorylated in
S. griseus, and this is a plausibleexplanation for the maintenance
of a multicopy of afsR. Thereduction of the growth rate may be
related to some un-known mechanism by which even an NH2-terminal
portionand a COOH-terminal portion still possess a
regulatoryfunction to some extent.
ACKNOWLEDGMENTS
This work was supported in part by a research grant from
theMinistry of Education, Science and Culture of Japan and by a
grantfrom the Kato Memorial Foundation for Bioscience
Researches.
REFERENCES1. Adamidis, T., P. Riggle, and W. Champness. 1990.
Mutations in
a new Streptomyces coelicolor locus which globally
blockantibiotic biosynthesis but not sporulation. J. Bacteriol.
172:2962-2969.
2. Aiba, H., F. Nakasai, S. Mizushima, and T. Mizuno.
1989.Phosphorylation of a bacterial activator protein, OmpR, by
aprotein kinase, EnvZ, results in stimulation of its
DNA-bindingability. J. Biochem. 106:5-7.
3. Anderson, D. J., and G. Globel. 1983. Immunoprecipitation
ofproteins from cell-free translations. Methods Enzymol.
96:111-120.
4. Bibb, M. J., K. F. Chater, and D. A. Hopwood. 1983.
Develop-ments in Streptomyces cloning, p. 53-82. In M. Inouye
(ed.),Experimental manipulation of gene expression. AcademicPress,
Inc., New York.
5. Bibb, M. J., and D. A. Hopwood. 1981. Genetic studies of
thefertility plasmid SCP2 and its SCP2* variants in
Streptomycescoelicolor A3(2). J. Gen. Microbiol. 126:427-442.
6. Burnett, W. N. 1981. "Western blotting": electrophoretic
trans-fer of proteins from sodium dodecyl sulfate-polyacrylamide
gelsto unmodified nitrocellulose and radiographic detection
withantibody and radioiodinated protein. Anal. Biochem.
112:195-
203.7. Champness, W., P. Riggle, and T. Adamidis. 1990. Loci
in-
volved in regulation of antibiotic synthesis. J. Cell.
Biochem.14A:88.
8. Cherry, J. R., T. R. Johnson, C. Dollard, J. R. Shuster,
andC. L. Denis. 1989. Cyclic AMP-dependent protein kinase
phos-phorylates and inactivates the yeast transcriptional
activatorADR1. Cell 56:409-419.
9. David, E. G. 1990. One-dimensional gel electrophoresis.
Meth-ods Enzymol. 182:195-203.
10. Davis, B. J. 1964. Disc electrophoresis-II: method and
applica-tion to human serum proteins. Ann. N.Y. Acad. Sci.
121:404-427.
11. Hara, O., S. Horinouchi, T. Uozumi, and T. Beppu.
1983.Genetic analysis of A-factor synthesis in Streptomyces
coeli-color A3(2) and Streptomyces griseus. J. Gen. Microbiol.
129:2939-2944.
12. Hess, F., K. Oosawa, N. Kaplan, and M. I. Simon.
1988.Phosphorylation of three proteins in the signaling pathway
ofbacterial chemotaxis. Cell 53:79-87.
13. Hess, F. J., R. B. Bourret, and M. I. Simon. 1988.
Histidinephosphorylation and phosphoryl group transfer in
bacterialchemotaxis. Nature (London) 336:139-143.
14. Hopwood, D. A., M. J. Bibb, K. F. Chater, T. Kieser, C.
J.Bruton, H. M. Kieser, D. J. Lydiate, C. P. Smith, J. M. Ward,and
H. Schrempf. 1985. Genetic manipulation of Streptomyces:a
laboratory manual. The John Innes Foundation, Norwich,United
Kingdom.
15. Hopwood, D. A., T. Kieser, H. M. Wright, and M. J. Bibb.
1983.Plasmids, recombination and chromosome mapping in
Strepto-myces lividans. J. Gen. Microbiol. 129:2257-2269.
16. Horinouchi, S., and T. Beppu. 1984. Production in large
quan-tities of actinorhodin and undecylprodigiosin induced by afsB
inStreptomyces lividans. Agric. Biol. Chem. 48:2131-2133.
17. Horinouchi, S., O. Hara, and T. Beppu. 1983. Cloning of
apleiotropic gene that positively controls biosynthesis of
A-fac-tor, actinorhodin, and prodigiosin in Streptomyces
coelicolorA3(2) and Streptomyces lividans. J. Bacteriol.
155:1238-1248.
18. Horinouchi, S., M. Kito, M. Nishiyama, K. Furuya, S.-K.
Hong,K. Miyake, and T. Beppu. 1990. Primary structure of AfsR,
aglobal regulatory protein for secondary metabolite formation
inStreptomyces coelicolor A3(2). Gene 95:49-56.
19. Horinouchi, S., Y. Kumada, and T. Beppu. 1984.
Unstablegenetic determinant of A-factor biosynthesis in
streptomycin-producing organisms: cloning and characterization. J.
Bacteriol.158:481-487.
20. Horinouchi, S., F. Malpartida, D. A. Hopwood, and T.
Beppu.1989. afsB stimulates transcription of the actinorhodin
biosyn-thetic pathway in Streptomyces coelicolor A3(2) and
Strepto-myces lividans. Mol. Gen. Genet. 215:355-357.
21. Horinouchi, S., H. Suzuki, and T. Beppu. 1986.
Nucleotidesequence of afsB, a pleiotropic gene involved in
secondarymetabolism in Streptomyces coelicolor A3(2) and
Streptomyceslividans. J. Bacteriol. 168:257-269.
22. Hunkapiller, M. W., E. Lujan, F. Ostrander, and L. E.
Hood.1983. Isolation of microgram quantities of protein from
poly-acrylamide gels for amino acid sequence analysis.
MethodsEnzymol. 91:227-236.
23. Jeffry, B. S., A. J. Ninfa, and A. M. Stock. 1989.
Proteinphosphorylation and regulation of adaptive responses in
bacte-ria. Microbiol. Rev. 53:450-490.
24. Kanamura, K., H. Aiba, S. Mizushima, and T. Mizuno.
1989.Signal transduction and osmoregulation in Escherichia coli.
J.Biol. Chem. 264:21633-21637.
25. Katz, E., C. J. Thompson, and D. A. Hopwood. 1983.
Cloningand expression of the tyrosinase gene from Streptomyces
anti-bioticus in Streptomyces lividans. J. Gen. Microbiol.
129:647-655.
26. Kessler, S. W. 1975. Rapid isolation of antigens from cells
witha staphylococcal protein A-antibody adsorbent: parameters ofthe
interaction of antibody-antigen complexes with protein A.
J.Immunol. 115:1617-1624.
27. Laemmli, U. K. 1970. Cleavage of structural proteins during
the
VOL. 173, 1991
on April 7, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
2318 HONG ET AL.
assembly of the head of bacteriophage T4. Nature
(London)227:680-685.
28. Lanka, E., C. Edelbluth, M. Schlicht, and H. Schuster.
1978.Escherichia coli dnaB protein. J. Biol. Chem.
253:5847-5851.
29. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.
Randall.1951. Protein measurement with the Folin phenol reagent.
J.Biol. Chem. 193:265-275.
30. Makino, K., H. Shinagawa, M. Amemura, T. Kawamoto, M.Yamada,
and A. Nakata. 1989. Signal transduction in the phos-phate regulon
of Escherichia coli involves phosphotransferbetween PhoR and PhoB
proteins. J. Mol. Biol. 210:551-559.
31. Ninfa, A. J., and B. Magasanik. 1986. Covalent modification
ofthe glnG product, NRI, by the glnL product, NRII, regulates
the transcription of the glnALG operon in Escherichia coli.Proc.
Natl. Acad. Sci. USA 83:5909-5913.
32. Ronson, C. W., B. T. Nixon, and F. M. Ausubel. 1987.
Con-served domains in bacterial regulatory proteins that respond
toenvironmental stimuli. Cell 49:579-581.
33. Stein, D., and S. N. Cohen. 1989. A cloned regulatory gene
ofStreptomyces lividans can suppress the pigment
deficiencyphenotype of different developmental mutants. J.
Bacteriol.171:2258-2261.
34. Weiss, V., and B. Magasanik. 1988. Phosphorylation of
nitrogenregulator 1 (NR1) of Escherichia coli. Proc. Natl. Acad.
Sci.USA 85:8919-8923.
J. BACTERIOL.
on April 7, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/