JOURNAL OF BACTERIOLOGY,
0021-9193/01/$04.00⫹0 DOI: 10.1128/JB.183.3.854–864.2001
Feb. 2001, p. 854–864 Vol. 183, No. 3
Copy igh © 2001, Ame ican Socie y o Mic obiology. All Righ s Rese ed.
Enhanced Symbio ic Pe o mance by Rhizobium opici
Glycogen Syn hase Mu an s
SILVIA MARROQUI
´,
1,2
ANGELES ZORREGUIETA,
1
CARMEN SANTAMARI
´A,
3
FRANCISCO TEMPRANO,
3
MARIO SOBERO
´N,
4
MANUEL MEGI
´AS,
2
AND J. ALLAN DOWNIE
1
*
John Innes Cen e, No wich NR4 7UH, Uni ed Kingdom
1
; Depa amen o de Mic obiologı´a y Pa asi ologı´a, Uni e sidad
de Se illa,
2
and CIFA “Las To es y Tomejil,” Alcala´ del Rı´o,
3
Se illa, Spain; and Depa amen o de Biologı´a
Molecula de Plan as, Ins i u o de Bio ecnologı´a, UNAM, Cue na aca, Mo elos, Me´xico
4
Recei ed 28 July 2000/Accep ed 7 No embe 2000
We isola ed a Tn5-induced Rhizobium opici mu an ha has enhanced capaci y o oxidize N,N-dime hyl-
p-phenylendiamine (DMPD) and he e o e has enhanced espi a ion ia cy och ome oxidase. The mu an had
inc eased le els o he cy och omes c
1
and CycM and a small inc ease in he amoun o cy och ome aa
3
. In plan
es s, he mu an inc eased he d y weigh o Phaseolus ulga is plan s by 20 o 38% compa ed wi h he con ol
s ain, hus showing signi ican ly enhanced symbio ic pe o mance. The p edic ed p oduc o he mu a ed gene
is homologous o glycogen syn hases om se e al bac e ia, and he mu an lacked glycogen. The DNA sequence
o he adjacen gene egion e ealed six genes p edic ed o encode p oduc s homologous o he ollowing gene
p oduc s om Esche ichia coli: glycogen phospho ylase (glgP), glycogen b anching enzyme (glgB), ADP glucose
py ophospho ylase (glgC), glycogen syn hase (glgA), phosphoglucomu ase (pgm), and glycogen deb anching
enzyme (glgX). All six genes a e ansc ibed in he same di ec ion, and analysis wi h lacZ gene usions sugges s
ha he i s i e genes a e o ganized in one ope on, al hough pgm appea s o ha e an addi ional p omo e ; glgX
is ansc ibed independen ly. Su p isingly, he glgA mu an had dec eased le els o high-molecula -weigh
exopolysaccha ide a e g ow h on glucose, bu le els we e no mal a e g ow h on galac ose. A dele ion mu an
was cons uc ed in o de o gene a e a nonpola mu a ion in glgA. This mu an had a pheno ype simila o ha
o he Tn5mu an , indica ing ha he enhanced espi a ion and symbio ic ni ogen ixa ion and dec eased
exopolysaccha ide we e due o mu a ion o glgA and no o a pola e ec on a downs eam gene.
Rhizobium opici induces ni ogen- ixing nodules on se e al
un ela ed opical legume plan s, including species o Phaseo-
lus, Leucaena, and Mac op ilium (30). Bac e ial espi a ion is
essen ial o symbio ic ni ogen ixa ion in wo ways: he ATP
necessa y o ni ogen ixa ion is de i ed om oxida i e phos-
pho yla ion, and espi a ion emo es oxygen, he eby p e en -
ing inac i a ion o ni ogenase by oxygen.
Rhizobia, like many o he bac e ia, possess b anched espi-
a o y chains wi h h ee o mo e e minal oxidases. The elec-
ons de i ed om di e en sou ces a e channeled o he qui-
none pool in he cy oplasmic memb ane and om he e a e
ans e ed di ec ly o quinol oxidases o , ia he cy och ome
bc
1
complex and cy och ome c, o cy och ome coxidases. The
espi a o y chains o B ady hizobium japonicum, Rhizobium le-
guminosa um, Rhizobium e li, and Azo hizobium caulinodans
ha e been s udied (1, 10, 18, 20, 28, 44). They all possess a
cy och ome coxidase o he aa
3
ype, which is a majo com-
ponen o he espi a o y chain in ae obiosis. This ype o
oxidase has also been desc ibed in R. opici (14). Al e na i e
cy och ome coxidases o quinol oxidases may also con ibu e
o ae obic espi a ion. In nodules, whe e oxygen le els a e low,
a cy och ome coxidase o he cbb
3
ype wi h a high a ini y o
oxygen acili a es espi a ion. The subuni s o his oxidase a e
encoded by he ixNOQP genes, which ha e been desc ibed in
mos hizobia (10, 18).
R. e li, like he ela ed species R. opici, can nodula e
Phaseolus beans. Mu an s o R. e li wi h inc eased espi a ion
(33, 44, 45, 46) we e isola ed on he basis o hei enhanced
capaci ies o oxidize N,N,N⬘,N⬘- e ame hyl-p-phenylendia-
mine (TMPD) by using he Nadi cy och ome oxidase es (29).
This es is based on he eac ion o TMPD and ␣-naph hol in
he p esence o cy och ome cand cy och ome coxidase o
p oduce indophenol blue. Rhizobium mu an s a ec ing o ma-
ion o cy och ome bc
1
, CycM, o cy och ome aa
3
a e Nadi
⫺
(10), whe eas mu an s wi h inc eased espi a ion ia cy o-
ch ome cs ain mo e s ongly (Nadi
⫹⫹
). Two R. e li Nadi
⫹⫹
mu an s had inc eased espi a ion due o induc ion o he
ixNOQP genes unde ee-li ing condi ions (34, 46). The
genes a ec ed in bo h mu an s a e linked o he pu ine bio-
syn he ic pa hway, and i was p oposed ha he in e media y
me aboli e 5-amino-4-imidazoleca boxamide ibonucleo ide
could ac as a nega i e e ec o o cy och ome cbb
3
p oduc ion
in R. e li (46). One o hese mu an s and wo uncha ac e ized
mu an s inc eased he ni ogen con en o Phaseolus ulga is
plan s by 22 o 25% compa ed wi h he wild- ype s ain (33, 44,
45). Mu an s o Sino hizobium melilo i wi h inc eased espi a-
ion and symbio ic pe o mance ha e also been desc ibed (54).
The aim o his s udy was o isola e and cha ac e ize R. opici
mu an s wi h enhanced espi a ion and symbio ic ni ogen ix-
a ion on P. ulga is plan s.
MATERIALS AND METHODS
Mic obiological echniques. The bac e ial s ains and plasmids used a e lis ed
in Table 1. R. opici and Ag obac e ium ume aciens s ains we e g own a 28°C
in TY medium (2) o Y minimal medium (43) supplemen ed wi h 10 mM
* Co esponding au ho . Mailing add ess: Depa men o Gene ics,
John Innes Cen e, Colney Lane, No wich NR4 7UH, Uni ed King-
dom. Phone: 44 1603 450000. Fax: 44 1603 450045. E-mail: allan
[email p o ec ed].
854
ammonium chlo ide and 0.2% (w / ol) succina e, glucose, galac ose, o manni ol.
Esche ichia coli s ains we e g own a 37°C in L medium (31), o which mal ose
(0.5% [w / ol]) was added o expe imen s in ol ing de ec ion o glycogen. An-
ibio ics we e added as app op ia e o he ollowing inal concen a ions (mic o-
g ams pe millili e ): ampicillin, 400; gen amicin (GEN), 10; kanamycin (KAN),
20; i ampin, 20; spec inomycin, 100; and e acycline, 10. Suc ose, when p esen ,
was added a 5% (w / ol).
Gene ic echniques. Plasmids we e ans e ed om E. coli DH5␣ o R. opici
by ipa en al ma ings wi h he helpe plasmid pRK2013 (11). R. opici CIAT899
was mu agenized wi h Tn5using pJB4JI (3), selec ing o mu an s on Y-succina e
medium supplemen ed wi h i ampin and KAN. Colonies we e sc eened using
he Nadi cy och ome oxidase es (29), which measu es cy och ome coxidase
ac i i y based on he eac ion o N,N-dime hyl-p-phenylendiamine o TMPD and
␣-naph hol, in he p esence o bac e ial cy och ome cand cy och ome coxidase,
o p oduce indophenol blue. Mu an s wi h enhanced ac i i y (Nadi
⫹⫹
) we e
isola ed as colonies s aining mo e s ongly. The app oxima ely 13-kb EcoRI
agmen ca ying he Tn5inse ion om he Nadi
⫹⫹
mu an A554 was cloned
in o he EcoRI si e in he mul iple cloning si e o pa ially diges ed pJQ200KS
(39). The esul ing plasmid (pIJ7796) was in oduced in o CIAT899, and selec-
ion was made o suc ose- esis an GEN-sensi i e ecombinan s. A639 was one
such ecombinan and was con i med by DNA hyb idiza ion o con ain Tn5in
he app op ia e loca ion. The glgA dele ion mu an A656 was gene a ed by a
ecip ocal c osso e , exchanging a dele ion de i a i e o glgA wi h he Tn5in
A554. Plasmid pIJ7883 ca ies glgA in a 2.2-kb ClaI-SacI inse . An in- ame
dele ion o glgA was cons uc ed by excising a 624-bp Ps I agmen o o m
pIJ7884, and hen he emaining 1.6-kb agmen was subcloned as a SalI-SacI
agmen in o pJQ200KS o o m pIJ7899. This plasmid was conjuga ed in o
A554 and suc ose- esis an , GEN and KAN-sensi i e ecombinan s we e iso-
la ed. One such isola e, A656, was con i med by DNA hyb idiza ion o ha e he
app op ia e pa e n o DNA agmen s. The usions o he glycogen me abolism
genes o he E. coli lacZ gene we e gene a ed by subcloning he ollowing DNA
agmen s in o pMP220 (47) in he co ec o ien a ion, esul ing in he plasmids
indica ed in pa en heses: 8-kb HindIII (glgP-lacZ; pIJ9160), 1.3-kb HindIII (glgB-
lacZ; pIJ9147), 2.4-kb HindIII (glgC-lacZ; pIJ9014), 0.7-kb EcoRI-XbaI(glgA-
lacZ; pIJ9016), 1.5-kb EcoRI-XbaI(pgm-lacZ; pIJ9015), 7.3-kb EcoRI-XbaI
(glgA::Tn5-pgm-lacZ; pIJ9034), 1.0-kb Ps I-EcoRI (glgA⌬Ps I-pgm-lacZ; pIJ9044),
and 0.8-kb XbaI-Ps I(glg-lacZ; pIJ9011) (Fig. 1B). ␤-Galac osidase ac i i y was
assayed essen ially as desc ibed by Rossen e al. (40) wi h cells g own in Y
medium supplemen ed wi h glucose o galac ose.
To con i m ha he pgm gene in pIJ7814 was exp essed, we ans e ed
pIJ7814 o s ain A5129 (a pgm mu an o A. ume aciens) (49) and obse ed
TABLE 1. Bac e ial s ains and plasmids
S ain o plasmid Geno ype o ele an cha ac e is ics Sou ce o e e ence
CIAT899 Wild- ype; Ri
Cm
Ap
30
A554 glgA1::Tn5; Nadi
⫹⫹
;Km
This wo k
A639 glgA1::Tn5ob ained by homologous ecombina ion This wo k
A656 glgA⌬Ps I This wo k
A. ume aciens
A348 Wild- ype; pTi
⫺
Cm
49
A5129 pgm::Tn5Km
49
E. coli
DH5␣F
⫺
supE44 ⌬lacU169 (␾80lacZ⌬M15) hsdR17 ecA1 41
endA1 gy A96 hi-1 elA1
LCB499 glgB de i a i e o PA601 7
RH97 glgC::Tn10 de i a i e o MC4100 R. Hengge-A onis
RH98 glgA::Tn10 de i a i e o MC4100 R. Hengge-A onis
Plasmids
pBluesc ip II KS pUC de i a i e; F1 eplica ion o igin; Ap
S a agene Cloning Sys ems
(La Jolla, Cali .)
pMP220 De i a i e o pKT214 ca ying lacZ wi hou a p omo e ; Tc
47
pJQ200KS sacB Lac Gm
39
pIJ1891 pLAFR3 de i a i e; Lac Tc
A. Ma idou
pJB4JI Suicide plasmid o Tn5mu agenesis o Rhizobium 3
pRK2013 Helpe plasmid ha p o ides plasmid ans e unc ions Km
11
pIJ7719 O e lapping cosmids om CIAT899 DNA lib a y which complemen A554 and A656 This wo k
pIJ7720
pIJ9158
pIJ7741 6.6-kb HindIII agmen om pIJ7720 ca ying glgA,pgm, and glgX in pBluesc ip II KS This wo k
pIJ7746 6.6-kb HindIII agmen om pIJ7720 in pIJ1891 complemen ing A554 and A656 This wo k
pIJ7796 glgA::Tn5in pJQ200KS This wo k
pIJ7814 1.6-kb EcoRI dele ion o pIJ7746 This wo k
pIJ7844 4.3-kb SacI agmen om pIJ7720 ca ying glgC and glgA in pBluesc ip II KS This wo k
pIJ7883 2.2-kb ClaI-SacI agmen ca ying glgA in o pBluesc ip II KS This wo k
pIJ7884 624-bp Ps I dele ion o glgA in pIJ7883 This wo k
pIJ7899 glgA⌬Ps I in pJQ200KS This wo k
pIJ7959 4.3-kb SacI subclone o pIJ7720 in pIJ1891 complemen ing A554 and A656 This wo k
pIJ9011 0.8-kb XbaI-Ps I agmen in pMP220; glgX-lacZ This wo k
pIJ9014 2.4-kb HindIII agmen in pMP220; glgC-lacZ This wo k
pIJ9015 1.5-kb EcoRI-XbaI agmen in pMP220; pgm-lacZ This wo k
pIJ9016 0.7-kb EcoRI-XbaI agmen in pMP220; glgA-lacZ This wo k
pIJ9034 pgm-lacZ bu ca ying Tn5o A554 This wo k
pIJ9044 pgm-lacZ bu ca ying Ps I dele ion This wo k
pIJ9147 1.3-kb HindIII agmen in pMP220; glgB-lacZ This wo k
pIJ9160 8-kb HindIII agmen in pMP220; glgP-lacZ This wo k
VOL. 183, 2001 R. TROPICI GLYCOGEN SYNTHASE MUTANTS 855
complemen a ion o he exopolysaccha ide-minus (EPS
⫺
) pheno ype on Y-man-
ni ol pla es.
DNA manipula ions. Va ious ope a ions we e ca ied ou acco ding o he
me hod o Samb ook e al. (41). DNA hyb idiza ions we e done a 65°C in 5⫻
SSC (1⫻SSC is 0.15 M NaCl plus 0.015 M sodium ci a e), and washes we e
done in 1⫻SSC–0.1% sodium dodecyl sul a e (SDS) a 65°C. Double-s anded
DNA was sequenced wi h a The mo Sequenase Dye Te mina o cycle sequenc-
ing ki (Ame sham) and he au oma ic sequence ABI377 (Pe kin-Elme ). The
glgC, glgA, pgm, and glgX genes we e sequenced on bo h s ands; only pa s o
glgP and glgB we e sequenced on bo h s ands. In addi ion o he uni e sal and
e e se p ime s, he ollowing p ime s we e used: 5⬘-GAAGTCAGATCCTGG
AAAACGGGAA-3⬘( o sequence one s and om he end o Tn5), 5⬘-GGAT
ACGTCGGATTTCATGCCCTG-3⬘( o sequence he Ps I dele ion used o gen-
e a e mu an A656), and 5⬘-CGTTGTCCTGGCAATAGGCA-3⬘( o comple e
he sequence o he glgX gene). The nucleo ide sequence was analyzed wi h he
Gene ics Compu e G oup e sion 8.1 package.
Cell ac iona ion, p o ein gel elec opho esis, heme s aining, and immuno-
s aining. R. opici cells g own on succina e we e ac iona ed in o memb anes
and soluble ac ions essen ially as desc ibed p e iously (9). P o ein concen a-
ions we e es ima ed by he me hod o B ad o d (5) using bo ine se um albumin
as a s anda d. The memb ane and soluble ac ions we e suspended in loading
bu e (124 mM T is [pH 7.0] 20% [ ol/ ol] glyce ol, 4.6% [w / ol] SDS), incu-
ba ed a 42°C o 20 min, sepa a ed by SDS-polyac ylamide gel elec opho esis
(PAGE), and ans e ed o a ni ocellulose il e . P o eins con aining co alen ly
bound heme we e s ained by chemiluminescence as desc ibed by Va gas e al.
(51).
Fo de ec ion o phosphoglucomu ase, cells we e g own o 48 h on glucose
minimal medium. The cells we e washed wi h 30 mM T is-HCl (pH 8.0)–3 mM
EDTA and esuspended in he same bu e con aining 20% (w / ol) suc ose, 200
␮g o lysozyme/ml, and 1 mM phenylme hanesul onyl luo ide. A e 1honice,
he cells we e cen i uged a 12,000 ⫻g o 30 min and esuspended in 50 mM
T is-HCl (pH 8.0), 10 mM MgCl
2
, 20% suc ose, 1 mM phenylme hanesul onyl
luo ide, and 10 mg o DNase I/ml. The cells we e b oken by wo passages
h ough a F ench p essu e cell, and he ex ac was cen i uged a 3,000 ⫻g o
30 min o emo e unb oken cells and hen a 100,000 ⫻g o 3h o emo e
memb anes. Soluble p o eins we e p ecipi a ed wi h 10% ichlo oace ic acid,
and he pelle was washed wi h ace one. Samples we e sepa a ed by SDS-PAGE,
ans e ed o ni ocellulose, incuba ed wi h a 1-in-500 dilu ion o an ise um o
he Ag obac e ium phosphoglucomu ase (48), and s ained wi h goa an i abbi
immunoglobulin conjuga ed o alkaline phospha ase (Sigma, S . Louis., Mo.).
Spec a and espi a o y ac i i ies. Fo spec a, R. opici cells we e washed in
esh Y-succina e medium, dilu ed o an op ical densi y o 0.05 a 540 nm, and
g own a 30°C o 12 h. The cells we e ha es ed by cen i uga ion, washed, and
suspended o 30% (w / ol) in 0.1 M phospha e bu e (pH 7.4) wi h 40% glyce ol.
Cy och ome spec a we e eco ded using an SLM Aminco Midan II spec opho-
ome e . The samples we e educed wi h a ew g anules o di hioni e o oxidized
wi h ammonium pe sul a e. Ca bon monoxide di e ence spec a ( educed ⫹CO
minus educed) we e ob ained by bubbling CO o 2 min h ough a di hioni e-
FIG. 1. (A) Physical and gene ic map o he R. opici glycogen egion and complemen a ion o he Nadi
2⫹
and low-EPS pheno ypes o mu an s
A554 and A656 by di e en R. opici cosmids. ⫹, complemen a ion; ⫺, no complemen a ion. The open a ows ep esen he glycogen me abolism
genes and he di ec ion o hei ansc ip ion. The b oken lines indica e ha he genes ha e no been sequenced comple ely. The posi ion o Tn5
in mu an A554 is indica ed by an in e ed iangle, and he agmen o glgA dele ed in mu an A656 is ep esen ed by a do ed segmen . The
do ed a ows indica e he pu a i e ansc ip ional uni s. B, BamHI; E, EcoRI; H, HindIII; P, Ps I; Sc, SacI. (B) T ansc ip ional usions o he
glycogen me abolism genes o E. coli lacZ. The di ec ion o ansc ip ion is indica ed by he a ows. The do ed segmen o pIJ9044 indica es he
egion dele ed. (C) Biochemical eac ions in ol ed in glycogen me abolism in A. ume aciens (adap ed om e e ence 49). The gene p oduc s
indica ed a e p oposed o ca alyze he co esponding eac ions.
856 MARROQUI
´ET AL. J. BACTERIOL.
educed cell suspension, and he spec a we e eco ded agains a educed sam-
ple. The spec a we e measu ed a oom empe a u e in 1.0-cm-diame e ligh
pa h cu e es. Fo he de e mina ion o oxygen up ake, he cells we e ha es ed
a e 48 h o g ow h a 28°C in Y-succina e medium and esuspended in 1 ml o
25 mM po assium phospha e bu e (pH 7.0). The oxygen up ake was measu ed
wi h a Hansa ech elec ode a 25°C a e addi ion o 2 mM TMPD and 10 mM
sodium asco ba e ( inal concen a ions).
Glycogen de e mina ion. The glycogen con en o R. opici s ains was de e -
mined as desc ibed by K isman (22) using R. opici cells g own o 3 days a 28°C
in 40 ml o Y-glucose medium con aining 3.5 ins ead o 10 mM NH
4
Cl. The
amoun o glycogen p esen was calcula ed om he abso bance a 540 nm in
ela ion o s anda d spec a eco ded using abbi glycogen (concen a ion ange,
20 o 300 ␮g/ml) as a s anda d.
Quan i a i e analysis o EPS. The s ains we e g own o 4 days in Y medium
supplemen ed wi h ei he glucose o galac ose a 28°C. The cul u es we e cen-
i uged a 10,000 ⫻g o 60 min, and he supe na an was collec ed, dilu ed 1:1
wi h dis illed wa e , and cen i uged again in o de o elimina e esidual cells.
The high-molecula -weigh EPS (HMW-EPS) was p ecipi a ed om he supe -
na an by adding sodium chlo ide ( o 0.3 M) and 2.5 olumes o e hanol. A e
16 h a 4°C, he p ecipi a ed HMW-EPS was spooled wi h a glass od and le o
d y a 37°C in a p eweighed pe i dish. The e hanol-p ecipi able ma e ial was
measu ed by weigh di e ence. The s uc u e o R. opici CIAT899 EPS was
p e iously de e mined (15).
Plan es s. P. ulga is c . Neg o Jamapa plan s (black beans) we e g own in a
g eenhouse in Leona d ja s wi h a ni ogen- ee medium o an a e age o 45
days, as desc ibed by Vincen (52). A leas ou eplica es we e used pe s ain.
Th ee pa ame e s we e measu ed: he d y weigh o he ae ial pa o he plan ,
he numbe o nodules o med, and he d y weigh o nodules pe plan . The
esul s we e analyzed s a is ically by analysis o a iance.
Nucluo ide sequence accession numbe . The DNA sequence was submi ed o
he EMBL da abase and has been assigned accession numbe AJ291603.
RESULTS
Isola ion and cha ac e iza ion o an R. opici Nadi
ⴙⴙ
mu-
an . Tn5-induced mu an s o R. opici CIAT899 we e
sc eened on succina e minimal medium using he Nadi es
(29). One mu an , A554, which s ained da ke blue (Nadi
⫹⫹
)
han he wild ype was chosen o de ailed s udy because i had
enhanced le els o symbio ic ni ogen ixa ion (see below).
This mu an had a highe a e o TMPD oxida ion han
CIAT899 (21 e sus 15 ng-a om o O
2
/min/mg [d y weigh ] o
cells). The Tn5and lanking DNA om A554 was cloned on an
EcoRI agmen in o he suicide ec o pJQ200KS (39) o o m
pIJ7796, and he Tn5was ecombined in o CIAT899. The
ecombinan , A639, had he same Nadi
⫹⫹
pheno ype as A554,
con i ming ha he Tn5inse ion was esponsible o he phe-
no ype. DNA hyb idiza ions wi h DNA om A554 and A639
demons a ed ha each ca ied only one copy o Tn5and ha
i was inse ed in he same posi ion in bo h s ains.
Th ee o e lapping cosmids we e isola ed om an R. opici
CIAT899 lib a y by colony hyb idiza ion, using DNA adjacen
o he Tn5inse ion as a p obe. The h ee cosmids, pIJ7719,
pIJ7720, and pIJ9158 (Fig. 1A), con ained a 6.6-kb HindIII
band which hyb idized o he same p obe. All h ee es o ed
he Nadi
⫹⫹
pheno ype o A554 and A639 o no mal (Fig. 1A)
bu had no e ec on he Nadi pheno ype o CIAT899. The
cloned 6.6-kb HindIII agmen (pIJ7746) also complemen ed
A554 and A639 o Nadi
⫹
(Fig. 1A).
Cy och ome composi ion o A554. The Nadi
⫹⫹
pheno ype o
A554 sugges ed ha he mu a ion may a ec he exp ession o
espi a o y chain componen s, so hese we e analyzed. Heme
s aining o soluble and memb ane ac ions o CIAT899 (Fig.
2) e ealed a majo soluble componen o app oxima ely 14
kDa, p obably co esponding o pe iplasmic cy och ome c, and
wo memb ane componen s o abou 20 and 31 kDa, p obably
co esponding o CycM and cy och ome c
1
. In o he wo k,
mu a ions a ec ing he genes encoding he cy och ome bc
1
complex ha e been shown o abolish he o ma ion o he
31-kDa componen (unpublished esul s). The heme s aining
o A554 is di e en om ha o CIAT899 in ha s aining o
he soluble 14-kDa componen is less in ense (Fig. 2A) and he
memb ane componen s o 31 (cy och ome c
1
) and 20 (CycM)
kDa s ained sligh ly mo e s ongly (Fig. 2B). This sugges s
highe le els o cy och ome c
1
and CycM in he mu an . Bo h
o hese cy och omes a e essen ial o a posi i e Nadi eac ion
in Rhizobium s ains (10). The inc eases in cy och ome c
1
and
CycM could accoun o he enhanced Nadi s aining in A554
compa ed wi h CIAT899.
The educed minus oxidized di e ence spec a (Fig. 3A) o
he wild- ype s ain e ealed peaks cha ac e is ic o cy o-
ch ome c(maximum a 553 nm), cy och ome b(maximum a
562 nm), and cy och ome aa
3
(maximum a 603 nm). A554 has
a simila spec um, al hough he a io o cy och ome b o c
seems o be highe han in CIAT899 and he e is a somewha
inc eased le el o cy och ome aa
3
. The ela i ely small change
in he cy och ome c egion o he spec um may be due o a
dec ease in abso p ion due o he soluble cy och ome c(Fig.
2A) being compensa ed o by inc eased le els o cy och ome
c
1
and CycM (Fig. 2B). The educed ⫹CO minus educed
di e ence spec a (Fig. 3B) con i med he p esence o cy o-
ch ome aa
3
(peak a 419 nm; oughs a 443 and 610 nm) and
e ealed spec oscopic ea u es cha ac e is ic o cy och ome o
o a high-spin cy och ome b(which a e indis inguishable by
his echnique) (peaks a 417 and 544 nm; oughs a 430 and
559 nm). The CO spec um o A554 is simila o ha o
CIAT899, excep ha i e eals a sligh ly inc eased amoun o
cy och ome aa
3
( oughs a 610 and 443 nm).
A554 has enhanced symbio ic pe o mance. P. ulga is c .
Neg o Jamapa beans we e g own unde ni ogen limi a ion
and inocula ed wi h he wild- ype CIAT899 o A554. Uninocu-
la ed con ol plan s g ew e y poo ly, and CIAT899 s ongly
s imula ed g ow h (Table 2). The a e age d y weigh s o plan s
inocula ed wi h he mu an A554 we e 38 and 20% g ea e
han hose obse ed wi h CIAT899 in expe imen s conduc ed
in wo sepa a e g owing seasons (Table 2). The yield di e ence
FIG. 2. C-heme s ains o soluble (A) and memb ane (B) p o eins
om R. opici CIAT899 (lanes 1) and A554 (lanes 2) a e SDS-
PAGE. Each lane was loaded wi h 50 (A) o 30 (B) ␮g o p o ein. The
molecula masses o he main bands a e indica ed and we e es ima ed
using p es ained molecula mass ma ke s (New England Biolabs) ha
do no show up on he heme-s ained gel.
VOL. 183, 2001 R. TROPICI GLYCOGEN SYNTHASE MUTANTS 857
be ween plan s inocula ed wi h A554 and CIAT899 in he i s
expe imen was signi ican a he 95% con idence le el, al-
hough he di e ence in he second expe imen was no signi -
ican a he 95% con idence le el. Howe e , using a wo-way
analysis o a iance combining he da a om bo h expe imen s,
A554 was ound o gi e a signi ican ly enhanced yield com-
pa ed wi h he wild ype (con idence le el, 98%).
The inc eased symbio ic pe o mance migh be due o en-
hanced ni ogen ixa ion by indi idual nodules o o inc eased
nodula ion. Those plan s wi h inc eased ni ogen ixa ion also
ha e an inc eased numbe o nodules and inc eased nodule
mass (Table 2). The same end is e iden om bo h expe i-
men s. The e o e, we conclude ha he enhanced symbio ic
pe o mance is p obably due o inc eased numbe s o ni ogen-
ixing nodules.
Cha ac e iza ion o he gene egion a ec ed in A554. The
DNA sequence o he 6.6-kb HindIII agmen ha comple-
men ed he Nadi
⫹⫹
pheno ype o A554 (Fig. 1A) e ealed he
p esence o ou open eading ames ansc ibed in he same
di ec ion. Th ee adjacen HindIII agmen s o 2.4, 1.3, and 8.0
kb (Fig. 1A) we e also subcloned and pa ially sequenced,
e ealing wo addi ional open eading ames also ansc ibed
in he same di ec ion. All six o he genes we e ound o encode
p o eins wi h sequences simila o hose o he p oduc s o E.
coli genes in ol ed in glycogen me abolism (37): glgP (encod-
ing glycogen phospho ylase), glgB (glycogen b anching en-
zyme), glgC (ADP glucose py ophospho ylase), glgA (glycogen
syn hase), pgm (phosphoglucomu ase), and glgX (glycogen de-
b anching enzyme). The oles o hese six p o eins in he syn-
hesis and deg ada ion o glycogen in E. coli a e illus a ed in
Fig. 1C. In E. coli and many o he bac e ia, glycogen syn hesis
p oceeds as ollows (37). Phosphoglucomu ase (Pgm) isome -
izes glucose-6-P o glucose-1-P, which is used o o m ADP-
glucose in a eac ion ca alyzed by ADP-glucose py ophos-
pho ylase (GlgC). This ac i a ed o m o glucose is
polyme ized ia ␣-1,4 linkages by glycogen syn hase (GlgA).
Subsequen ly, he b anching enzyme (GlgB) o ms ␣-1,6-gly-
cosidic linkages. Glycogen is deg aded by deb anching enzyme
(GlgX), which emo es he ␣-1, 6 linkages, and by glycogen
phospho ylase (GlgP), which deg ades he ␣-1,4 linkages, gi -
ing ise o glucose-1-P. The o ganiza ion o he glycogen me-
abolism genes in R. opici is simila o ha epo ed o A.
ume aciens (48, 50), al hough he glgX gene was no desc ibed
in A. ume aciens.
(i) glgP.The R. opici glgP-like gene was only pa ially se-
quenced. A 621-bp agmen showed 91% simila i y a he
p o ein le el wi h A. ume aciens GlgP. The 3⬘end o he R.
opici glgP gene was also sequenced (77% simila i y a he
p o ein le el wi h A. ume aciens GlgP in a 294-bp agmen ),
displaying an o e lap be ween he p edic ed glgP e mina ion
codon (TGA) and he glgB ini ia ion codon (ATGA), sugges -
ing ansla ional coupling (21).
(ii) glgB.The 5⬘and 3⬘ends o R. opici glgB we e se-
quenced, showing high simila i y wi h A. ume aciens glgB (66%
in 285 bp and 89% in 405 bp a he p o ein le el, espec i ely),
which was p oposed o encode glycogen b anching enzyme
based on sequence simila i y wi h he E. coli gene (48).
FIG. 3. (A) Reduced minus oxidized di e ence spec a o R. opici
wild- ype (CIAT899) and glycogen syn hase mu an (A554) s ains. (B)
Reduced ⫹CO minus educed di e ence spec a o s ains CIAT899
and A554. The numbe s ep esen he wa eleng hs (in nanome e s) a
which he main peaks and oughs a e de ec ed. The cells we e g own
in Y-succina e minimal medium o 12 h. The p o ein concen a ions
we e as ollows: CIAT899, 15.2 mg/ml; A554, 17.0 mg/ml.
TABLE 2. Symbio ic pheno ype o R. opici wild ype and glgA mu an s ains on P. ulga is c . Neg o Jamapa plan s
a
S ain Exp 1 Exp 2
Plan d y w (g) No. o nodules Nodule d y w (mg) Plan d y w (g) No. o nodules Nodule d y w (mg)
Uninocula ed con ol 0.355 a 0 d 0 0.563 i 0 l 0 o
CIAT899 3.181 b 366 e 239 g 4.335 j 557 m 272 p
A554 4.394 c 416 e 290 h 5.219 jk 764 n 439 q
A656 5.403 k 689 mn 401 q
a
Values ollowed by di e en le e s a e s a is ically signi ican ly di e en om each o he (P⬍0.05) wi hin each expe imen based on analysis o a iance.
858 MARROQUI
´ET AL. J. BACTERIOL.

(iii) glgC.The glgC gene is 26 bp downs eam o glgB and is
p eceded by an A/G- ich egion, pa o which could cons i u e
a ibosome-binding si e. The deduced p o ein (420 esidues) is
homologous o GlgC om A. ume aciens and E. coli (95 and
73% simila i y, espec i ely).
(i ) glgA.The glgA gene is 3 bp downs eam o glgC and does
no ha e an ob ious ibosome-binding si e. The deduced p o-
ein (480 esidues) is 90 and 66% homologous wi h GlgA om
A. ume aciens and E. coli, espec i ely. In E. coli GlgA, wo
impo an si es ha e been desc ibed, and hey a e conse ed in
he R. opici glycogen syn hase: he lysine a posi ion 15, which
o ms pa o he mo i KXGG (whe e X ep esen s any amino
acid) and is in ol ed in he binding o he ADP-glucose sub-
s a e (12), and he a ginine a posi ion 277 (lysine in he E. coli
p o ein), which cons i u es pa o he p oposed ac i e si e (13).
DNA sequencing o a 3.5-kb EcoRI-BamHI agmen ca ying
pa o he Tn5cloned om A554 e ealed ha he ansposon
is inse ed wi hin glgA a a posi ion co esponding o amino
acid esidue 304. This mu a ion was called glgA1::Tn5.
( ) pgm.A gene encoding a phosphoglucomu ase homolog is
included in he glycogen egion o R. opici, as is ound in A.
ume aciens (50), bu in E. coli i is elsewhe e in he genome
(27). The R. opici pgm gene homolog is 4 bp downs eam o
he glgA s op codon and is p eceded by a possible ibosome-
binding si e, GAGAGG, 11 bp om he ATG. The deduced
Pgm p o ein (542 esidues) is 88 and 50% simila o he A.
ume aciens and E. coli Pgms, espec i ely.
Downs eam o pgm in R. opici is a 211-bp noncoding egion
ha has wo in e ed- epea segmen s spanning 27 and 26 bp,
espec i ely, which could cons i u e a ho-independen ansc ip-
ion e mina o . A simila egion was iden i ied elsewhe e in R.
opici (sequence U47030 om he EMBL gene bank) (36).
FIG. 4. P o ein sequence compa ison o he glycogen deb anching enzymes (GlgX) o R. opici and E. coli (simila i y, 61%; iden i y, 41%). The
as e isks ep esen iden ical amino acid esidues. The do s indica e conse ed subs i u ions. Impo an esidues conse ed be ween R. opici and
E. coli p o eins a e ma ked wi h lines.
VOL. 183, 2001 R. TROPICI GLYCOGEN SYNTHASE MUTANTS 859
( i) glgX.The e is an R. opici glgX homolog 211 bp down-
s eam o pgm, and i is p eceded by a possible ibosome-
binding si e, GGAAAG, 10 bp om he p edic ed ATG. The
deduced p o ein (656 esidues) is homologous o he p edic ed
deb anching enzymes o many bac e ia, including E. coli (656
amino acids; 61% simila i y) (53) (Fig. 4). R. opici GlgX also
shows 56% simila i y wi h he Fla obac e ium sp. isoamylase
enzyme, which deg ades ␣-1,6 glycosidic bonds (23). This sim-
ila i y is p ima ily a ound he ou domains (indica ed in Fig.
4) common o amyloly ic enzymes (19).
The glgA mu an lacks glycogen. The glycogen con en s o
s ains CIAT899, A554 (glgA1::Tn5), and he complemen ed
s ain A554/pIJ7720 we e analyzed. CIAT899 and he comple-
men ed s ain had simila le els o glycogen (1.7 and 1.6 ng/10
6
CFU, espec i ely), whe eas A554 had essen ially no glycogen
(⬍0.01 ng/10
6
CFU). This con i ms ha he ansposon in-
se ed in he glycogen syn hase gene o A554 p e en s glycogen
o ma ion and ha he mu a ion is complemen ed by cosmid
pIJ7720 ca ying he glgA egion.
The cloned gene egion om R. opici was es ed o i s
abili y o complemen E. coli mu an s wi h a ec ed glycogen
syn hesis. Colonies o glycogen-de icien mu an s o E. coli can
be dis inguished om he wild ype by iodine s aining (16).
This assay was used o analyze complemen a ion o he E. coli
glgA, glgB, and glgC mu an s (RH98, LCB499, and RH97, e-
spec i ely [Table 1]) by R. opici plasmids ca ying glg genes.
pIJ7741 pa ially complemen ed RH98, pIJ7844 comple-
men ed RH98 and pa ially complemen ed RH97, and pIJ9158
complemen ed LCB499. We conclude ha he glycogen p o-
duc ion o hese mu an s was ully o pa ially es o ed by he
R. opici glgA, glgB, and glgC genes.
Analysis o ope on s uc u e using lacZ usions. T ansc ip-
ional usions o all six genes we e cons uc ed using he lacZ
epo e plasmid pMP220 (47) (Fig. 1B). The esul ing plas-
mids we e in oduced in o CIAT899, and he ␤-galac osidase
ac i i y was measu ed (Table 3). Signi ican le els o ac i i y
we e de ec ed o he glgP-lacZ (pIJ9160) and pgm-lacZ
(pIJ9015) gene usions, whe eas he glgB-lacZ (pIJ9147), glgC-
lacZ (pIJ9014), and glgA-lacZ (pIJ9016) usions had e y low
le els o ac i i y. The glgX-lacZ usion (pIJ9011) had ␤-galac-
osidase ac i i y, bu he le el obse ed was low. These esul s
indica e ha he e is an ope on composed o glgP, glgB, glgC,
and glgA. The pgm gene appea s o ha e a sepa a e p omo e ,
al hough he p esence o only 4 bp be ween he p edic ed
coding egions o glgA and pgm sugges s ha pgm is also ex-
p essed wi hin he glgPBCA ope on. Such a po en ial dual
con ol o pgm exp ession seems easonable in iew o he ac
ha pgm has an impo an ole in me abolism independen o
glycogen biosyn hesis (Fig. 1C). In A. ume aciens, he glgP,
glgB, glgC, glgA, and pgm genes a e ansc ibed om a p o-
mo e loca ed ups eam o glgP (48). The A. ume aciens pgm
gene can also be ansc ibed as a sho e p oduc om a
p omo e immedia ely ups eam o an in e nal ansla ion ini-
ia ion codon wi hin he pgm gene o yield a sho e Pgm
p o ein (48). DNA sequence compa isons e ealed an in- ame
ATG wi hin he R. opici pgm gene in a con ex almos iden-
ical o he sequence ound in A. ume aciens. The e o e, pgm
in R. opici may be ansc ibed independen ly, as is seen in A.
ume aciens.
The glgX gene appea s o be in a sepa a e ansc ip ional
uni , based on he obse a ions ha he e is a po en ial an-
sc ip ion e mina o be ween pgm and glgX and ha he glgX-
lacZ usion is exp essed.
Analysis o a nonpola glgA dele ion mu an . To de e mine
i he pleio opic pheno ypes o he glgA mu an A554 a e due
o pola i y o he Tn5, an in- ame dele ion mu a ion in glgA
was cons uc ed in i o and ecombined in o he genome o R.
opici o o m A656. This mu an was simila o A554 in ha
i had a Nadi
⫹⫹
pheno ype on Y-succina e pla es and had
inc eased le els o TMPD oxida ion (23 ng-a om o O
2
/min/mg
[d y weigh ]) compa ed wi h CIAT899 (15 ng-a om o O
2
/
min/mg [d y weigh ]). Heme s aining e ealed ha he cy o-
ch ome ccon en was essen ially he same as ha seen wi h
A554, in ha he con en o cy och ome c
1
and CycM was
inc eased and he soluble cy och ome ccon en was dec eased
(da a no shown).
The g ow h o P. ulga is plan s inocula ed wi h he glgA
dele ion mu an A656 was de e mined in pa allel wi h one o
he es s o he glgA::Tn5mu an . As shown in Table 2, A656
pe o med signi ican ly be e (25%; P⬍0.05) han he con ol
s ain, CIAT899. I beha ed indis inguishably om A554, and
i he esul s om he h ee independen es s o he wo glgA
mu an s a e pooled in a wo-way analysis o a iance, i is
e iden ha mu a ion o glgA induces signi ican ly inc eased
g ow h (P⬍0.02) compa ed wi h CIAT899.
The gene downs eam o glgA is pgm. The A. ume aciens
phosphoglucomu ase has been pu i ied, and an an ise um has
been p epa ed (48). To con i m ha he dele ion mu a ion in
glgA does no ha e a pola e ec on pgm, we used he an i-
se um p oduced agains he A. ume aciens Pgm o de e mine
he amoun o Pgm p o ein in A656 and he wild ype,
CIAT899 (Fig. 5). I is e iden ha he le el o s aining in he
mu an is indis inguishable om ha seen wi h he wild ype,
con i ming ha he dele ion mu a ion does no cause a pola
e ec on pgm exp ession. Simila esul s we e seen in h ee
independen analyses. In A. ume aciens, wo di e en -size
Pgm p o eins a e p oduced as a esul o di e en ansc ip ion
and ansla ion s a si es in he pgm gene (48). In wild- ype R.
opici, he e appea o be wo bands ecognized by he an i-
se um, and hese a e also p esen in he mu an (Fig. 5A). The
es ima ed size o he p oduc s is abou 60 kDa, which is close
o he molecula mass (58.4 kDa) p edic ed om he DNA
sequence. The di e ence be ween he wo p oduc s is es i-
ma ed o be abou 2 kDa. These expe imen s do no dis inguish
TABLE 3. Exp ession o glg- and pgm-lacZ ansc ip ional usions
in R. opici
Plasmid ␤-Galac osidase ac i i y
a
Glucose Galac ose
pMP220 150 ⫾1 165 ⫾20
pIJ9160 (glgP-lacZ) 1,425 ⫾96 746 ⫾54
pIJ9147 (glgB-lacZ)54⫾2ND
pIJ9014 (glgC-lacZ)97⫾2ND
pIJ9016 (glgA-lacZ)60⫾2ND
pIJ9015 (pgm-lacZ) 1,787 ⫾156 1,303 ⫾30
pIJ9011 (glgX-lacZ) 283 ⫾49 172 ⫾13
a
Ac i i y is exp essed in Mille uni s ⫾s anda d e o . The cells we e g own
in minimal medium supplemen ed wi h glucose o galac ose o 24 h. The esul s
a e a e ages o a leas h ee expe imen s. ND, no de e mined.
860 MARROQUI
´ET AL. J. BACTERIOL.
be ween wo dis inc ansla ion p oduc s and p ocessing o a
single p oduc .
glgA mu an s ha e educed EPS con en . The glycogen syn-
hase mu an A554 was o iginally iden i ied ollowing g ow h
on succina e minimal medium pla es, on which i g ows no -
mally. I also o ms no mal-size colonies on galac ose o man-
ni ol minimal medium o on comple e medium (TY). How-
e e , on glucose o mal ose he colony size was educed
signi ican ly and he colonies appea ed o ha e less EPS. In
liquid glucose minimal medium, he mu an s had a somewha
ex ended lag phase, bu he g ow h a es we e no mal.
Quan i a i e analysis o HMW-EPS p oduced in liquid cul-
u es (Table 4) con i med ha A554 p oduces less HMW-EPS
han CIAT899 a e g ow h on glucose bu p oduces simila
amoun s o HMW-EPS ollowing g ow h on galac ose. The
mu a ion is complemen ed by pIJ9158, which es o es no mal
le els o EPS p oduc ion (Table 4). The ela i e amoun s o
HMW-EPS shown in Table 4 a e based on he yield o EPS pe
millig am (d y weigh ) o cells; simila educ ions in he
HMW-EPS yield we e seen in o he p epa a ions o he mu-
an g own on glucose i he yield was calcula ed ela i e o he
numbe o CFU (da a no shown). I seemed unlikely ha
mu a ion o glgA would di ec ly cause a dec ease in HMW-EPS
le els. A mo e likely explana ion was ha he Tn5inse ion in
A554 migh ha e a pola e ec on he downs eam pgm gene,
which is in ol ed in he o ma ion o glucose-1-P, which is a
p ecu so o bo h glycogen and HMW-EPS (Fig. 1C). To es
his, we analyzed HMW-EPS p oduc ion by he nonpola glgA
dele ion mu an A656. Su p isingly, A656 has a pheno ype
simila o ha o A554, p oducing a low le el o HMW-EPS
ollowing g ow h on glucose bu no galac ose (Table 4). Col-
onies o A656 we e, like A554, no mal on manni ol and galac-
ose minimal medium bu we e smalle on glucose minimal
medium. This sugges s ha he glgA mu a ion i sel in luences
he le el o HMW-EPS p oduced on glucose medium.
To u he con i m ha i is loss o glgA ha causes a de-
c ease in HMW-EPS p oduc ion ( a he han pola e ec s o
he mu a ions on pgm), we in oduced in o he glgA mu an s
di e en plasmids ca ying glgA o pgm (Fig. 1A). On glucose
pla es, a plasmid (pIJ7959) ca ying glgA (bu no pgm) e-
s o ed he sizes o colonies o no mal. Con e sely, pIJ7814,
which exp esses pgm bu lacks glgA, had no e ec on A554 o
A656 colony size on glucose medium. These complemen a ion
esul s ha e he same pa e n as hose ob ained using he Nadi
es (Fig. 1A), demons a ing ha hese wo pheno ypes a e
co ela ed. The complemen a ion o HMW-EPS p oduc ion
was con i med by quan i a i e analysis o HMW-EPS p oduced
in liquid cul u e using he wo glgA mu an s A554 and A656
ca ying he a ious plasmids. Those s ains wi h es o ed
EPS
⫹
colony mo phology (Fig. 1A) all o med amoun s o
HMW-EPS simila o hose obse ed when pIJ9158 was
p esen (da a no shown). The le el o HMW-EPS p oduced by
he glgA mu an s ca ying pIJ7814 was simila o ha seen
(Table 4) wi h A554 o A656 lacking plasmid (da a no shown).
We also measu ed he le el o pgm-lacZ exp ession in plas-
mid cons uc s ca ying he glgA1::Tn5o glgA dele ion
(pIJ9034 and pIJ9044, espec i ely [Fig. 1B]). In bo h cases,
he le el o pgm exp ession was simila o ha obse ed p e-
iously wi h he pgm-lacZ usion on pIJ9015 (Table 3).
On he basis o all o hese esul s, we conclude ha mu a-
ion o he glgA gene in some way dec eases EPS p oduc ion in
R. opici g own in glucose and ha his is unlikely o be due o
a pola e ec on pgm.
DISCUSSION
I had been epo ed p e iously ha some hizobial mu an s
which had inc eased espi a o y capaci ies had enhanced sym-
bio ic ni ogen ixa ion (33, 44, 45), and i was ou aim o
iden i y mu an s o R. opici ha had inc eased espi a ion and
hence symbio ic ni ogen ixa ion. The mu an s ain A554 was
in es iga ed in de ail because i ul illed bo h o hese c i e ia,
and i was somewha unexpec ed o ind ha he mu a ion
FIG. 5. Immunos aining o phosphoglucomu ase in A656
(glgA⌬Ps I). Soluble p o eins om he glgA dele ion mu an A656 and
he wild ype, (w ), CIAT899, we e sepa a ed by SDS-PAGE and
ei he s ained wi h Coomassie blue (B) o immunos ained wi h an i-
se um o phosphoglucomu ase (A). The same amoun o p o ein (10
␮g) was loaded in each lane. The sizes (in kilodal ons) o he
p es ained s anda ds (s ) a e indica ed o he igh o he gel.
TABLE 4. EPS p oduced by R. opici wild- ype and glgA mu an
s ains
a
S ain
HMW-EPS (mg/mg [d y w ] o
cells ⫾SE)
Glucose Galac ose
CIAT899 1.8 ⫾0.3 1.8 ⫾0.3
CIAT899/pIJ9158 1.6 ⫾0.2 1.9 ⫾0.3
A554 (glgA::Tn5) 0.7 ⫾0.1 1.6 ⫾0.3
A554/pIJ9158 1.5 ⫾0.2 1.8 ⫾0.3
A656 (glgA⌬Ps I) 0.5 ⫾0.1 2.0 ⫾0.3
A656/pIJ9158 2.2 ⫾0.3 2.0 ⫾0.4
a
Expe imen s we e epea ed a leas h ee imes. On each occasion, he
HMW-EPS p oduc ion by he glgA mu an s g own on glucose was less han 40%
ha o con ols, and no signi ican di e ence was seen wi h galac ose-g own cells.
VOL. 183, 2001 R. TROPICI GLYCOGEN SYNTHASE MUTANTS 861
a ec ed glycogen syn hesis. Mu a ion o he glycogen syn hase
(glgA) gene o R. opici induces pleio opic e ec s in addi ion
o he expec ed block o glycogen o ma ion. I is no imme-
dia ely ob ious why mu a ion o glgA should also lead o (i)
al e a ion in cy och omes, p esumably causing inc eased abili y
o oxidize TMPD; (ii) dec eased EPS p oduc ion du ing
g ow h on glucose (bu no galac ose o manni ol); and (iii)
inc eased nodula ion, appa en ly esul ing in enhanced symbi-
o ic pe o mance wi h he Phaseolus bean. Indeed, we canno
be su e ha hese pheno ypes a e all di ec consequences o
he inabili y o o m glycogen o which, i any, o hese phe-
no ypes causes he o he . Thus, o example, we do no know
i enhanced symbio ic pe o mance is due o he absence o
glycogen, inc eased espi a o y po en ial, o an al e a ion in
he le el o HMW-EPS. I may be o pa icula signi icance
ha an R. e li mu an de ec i e o s o age o he o he majo
sou ce o s o ed ca bon (poly-␤-hyd oxybu y a e) causes en-
hanced symbio ic pe o mance (8).
In E. coli, glycogen me abolism is highly egula ed a a me -
abolic le el and is also unde complex gene ic con ol (37).
Se e al me aboli es ha e posi i e o nega i e con ol o e gly-
cogen gene exp ession. Cyclic AMP (cAMP) ecep o p o ein-
cAMP and ppGpp a e he main posi i e egula o s o he
exp ession o he glgCAP ope on in E. coli. cAMP ecep o
p o ein-cAMP also s imula es exp ession o ano he gene in-
ol ed in glycogen biosyn hesis, glgS, whose unc ion is un-
known (17). The s a iona y-phase sigma ac o ␴
s
(24) has a
posi i e con ol o e glgS. On he o he hand, h ee nega i e
e ec o s ha e been desc ibed: one ha ac s in cis (glgR) o e
he glgCAP ope on, one ha ac s in ans (glgQ) o e he glgBX
and glgCAP genes (37), and he ca bon s o age egula o cs A,
which nega i ely a ec s all glycogen me abolism genes (53) by
speci ically des abilizing mRNA (26).
I is possible ha mu a ing glgA in R. opici may esul in a
change in he le els o glucose phospha es and/o suga nucle-
o ides, such as ADP-glucose, and ha his may a ec aspec s o
me abolism o he han glycogen syn hesis. In e es ingly, mu a-
ions o he S. melilo i cya3 gene encoding an adenyl cyclase-
like p o ein also enhanced symbio ic e ec i eness (42). In E.
coli, in acellula UDP-glucose has egula o y e ec s ia RpoS
(4), and i is possible ha some such egula o y e ec could
occu in R. opici. Accumula ion o suga nucleo ides o glu-
cose phospha es migh accoun o he pleio opic pheno ypes,
bu his needs o be es ed expe imen ally. One way o es he
e ec s o accumula ion o ADP-glucose would be o gene a e
a nonpola glgC mu an . This should be de ec i e o glycogen
syn hesis bu would be p edic ed no o accumula e ADP-
glucose (Fig. 1C) and so could be used o dis inguish he e ec s
o he absence o glycogen om possible accumula ion o
ADP-glucose. Glycogen and EPS biosyn hesis a e connec ed a
he le el o glucose-1-P (Fig. 1C), and he glucose-dependen
ep ession o HMW-EPS biosyn hesis in glgA mu an s may be
a consequence o inc eased le els o glucose phospha es, which
could ha e some kind o inhibi o y eedback e ec (di ec ly o
indi ec ly) on enzymes o HMW-EPS biosyn hesis. A key en-
zyme common o bo h pa hways is phosphoglucomu ase; when
a mu a ion in his gene was i s iden i ied in A. ume aciens
and S. melilo i, he gene was o iginally called exoC (6) because
o he s ong e ec on EPS o ma ion. La e , he exoC gene
was enamed pgm (48) when i s unc ion was mo e clea ly
unde s ood. The obse a ion ha pgm is immedia ely down-
s eam o glgA led us o suspec ha he dec eased le el o
HMW-EPS in he glgA Tn5mu an was due o a pola e ec on
pgm, bu we ound no e idence o such an e ec .
The obse a ion ha he HMW-EPS- educed pheno ype o
glgA mu an s is speci ic o g ow h on glucose bu no o he
suga s, such as galac ose, also a gues agains a simple pola
e ec on pgm.Apgm mu an o Ag oba e ium lacks EPS on
bo h galac ose and glucose media, due o i s inabili y o o m
UDP-glucose (49), and i is likely ha a simila si ua ion would
occu in Rhizobium. The no mal le el o p oduc ion o HMW-
EPS by he R. opici glgA mu an on galac ose he e o e im-
plies ha pgm is exp essed in he glgA mu an . Ou analysis o
gene exp ession indica ed ha he pgm gene is exp essed unde
a p omo e sepa a e om ha o glgA, al hough i s loca ion
immedia ely downs eam o glgA would sugges i is also an-
sc ibed along wi h glgA.
A pu a i e pgm mu an o R. opici CIAT899 was epo ed
p e iously (32). This mu an had educed le els o EPS (al-
hough he educ ion was no as g ea as in he R. opici exo
mu an s cha ac e ized in he same wo k). The mu an was
no mal wi h espec o mo ili y and a symbio ic pheno ype on
bean. This is di e en om S. melilo i o A. ume aciens pgm
(exoC) mu an s, which lack EPS, ha e al e ed LPS, a e non
mo ile, and a e de ec i e o symbiosis o a i ulen , espec-
i ely (6, 25). The epo ed pheno ype o he R. opici exoC
mu an is simila o ha o he glycogen syn hase mu an de-
sc ibed he e. The inse ion o Tn5in he pgm gene was no
demons a ed (32), and i was assumed o be a pgm (exoC)
mu an based on complemen a ion o an S. melilo i exoC mu-
an by a cosmid ha ca ied o he genes (32). We suspec ed
ha his p e iously desc ibed mu a ion migh be in glgA and
ied o cons uc a pgm mu an o R. opici. A spec inomycin-
esis an in e poson (38) was cloned a wo di e en si es o
pgm in wo di e en cons uc s. We in eg a ed he pgm::⍀
alleles on a sacB-based suicide-selec ion plasmid in o he ge-
nome bu could no gene a e a second ecombina ion e en o
make a pgm mu an unless a complemen ing clone ca ying
pgm was in oduced. This sugges s ha he pgm gene may be
essen ial o g ow h in R. opici. Du ing he cou se o ou
wo k, we had ied o in oduce he cloned glgA genes om R.
opici o A. ume aciens in o a pgm mu an o A. ume aciens,
and in bo h cases we ailed o ob ain ansconjugan s wi h he
pgm mu an e en hough a es o conjuga ion in o he wild-
ype s ain we e no mal. This sugges s ha in an A. ume aciens
pgm mu an backg ound, inc eased exp ession o glgA may be
le hal. Possibly ou inabili y o gene a e a pgm mu an o R.
opici is a ela ed e ec .
The inc ease in plan g ow h s imula ed by R. opici glgA
mu an s could in p inciple be due o enhanced e iciency o
ni ogen ixa ion by bac e oids, inc eased nodula ion, o bo h.
The p ima y e ec seems o be due o enhanced nodula ion.
The inc eased nodula ion seems unlikely o be due o al e ed
espi a ion, bu his canno be o mally uled ou . A dec ease
in EPS o ma ion could in p inciple a ec nodula ion, since
EPS plays a signaling ole du ing nodula ion (35), and indeed,
mu a ions a ec ing he le els o HMW-EPS in S. melilo i also
enhanced symbio ic pe o mance (42). Al e na i ely, i is pos-
sible ha he inabili y o s o e glycogen in some way causes
mo e e icien in ec ion. Howe e , in gene al, he plan ends
862 MARROQUI
´ET AL. J. BACTERIOL.