Vol. 149, No. 1JOURNAL OF BACTERIOLOGY, Jan. 1982, p. 65-780021-9193/82/010065-14$02.WO

Biochemical Diversity for Biosynthesis of Aromatic AminoAcids Among the Cyanobacteria

GERALDINE C. HALL,2 MARYANN B. FLICK,2 ROBERT L. GHERNA,3 AND ROY A. JENSEN'*Centerfor Somatic-cell Genetics and Biochemistry' and Department ofBiological Sciences,2 State University

ofNew York at Binghamton, Binghamton, New York 13901, and American Type Culture Collection,3Rockville, Maryland 20852

Received 6 April 1981/Accepted 28 August 1981

We examined the enzymology and regulatory patterns of the aromatic aminoacid pathway in 48 strains of cyanobacteria including representatives from each ofthe five major groupings. Extensive diversity was found in allosteric inhibitionpatterns of 3-deoxy-D-arabinoheptulosonate 7-phosphate synthase, not only be-tween the major groupings but also within several of the generic groupings.Unimetabolite inhibition by phenylalanine occurred in approximately half of thestrains examined; in the other strains unimetabolite inhibition by tyrosine andcumulative, concerted, and additive patterns were found. The additive patternssuggest the presence of regulatory isozymes. Even though both arogenate andprephenate dehydrogenase activities were found in some strains, it seems clearthat the arogenate pathway to tyrosine is a common trait that has been highlyconserved among cyanobacteria. No arogenate dehydratase activities were found.In general, prephenate dehydratase activities were activated by tyrosine andinhibited by phenylalanine. Chorismate mutase, arogenate dehydrogenase, andshikimate dehydrogenase were nearly always unregulated. Most strains preferredNADP as the cofactor for the dehydrogenase activities. The diversity in theallosteric inhibition patterns for 3-deoxy-D-arabinoheptulosonate 7-phosphatesynthase, cofactor specificities, and the presence or absence of prephenatedehydrogenase activity allowed the separation of subgroupings within several ofthe form genera, namely, Synechococcus, Synechocystis, Anabaena, Nostoc, andCalothrix.

The cyanobacteria (blue-green algae) com-prise a very large, diverse group of procaryoticorganisms which are of tremendous importancebecause of their photosynthetic and nitrogen-fixing capabilities. Despite this importance,however, the cyanobacteria are poorly under-stood in terms of their physiology, development,and genetics. Moreover, it is not known to whatextent enzymological diversity exists among thebiosynthetic pathways of these photoautotro-phic organisms.The term "enzymological patterning" in-

cludes differences in allosteric control patternsfor biochemical pathways, differences in cofac-tor specificities, and pathway multiplicity, i.e.,different biochemical routes to the same endproduct. Pathway multiplicity among microor-ganisms has been recently reviewed and occursin the biosynthesis of arginine, isoleucine, ly-sine, methionine, and serine (21). Different met-abolic pathways for the biosynthesis of phenyl-alanine and tyrosine using the newly discoveredintermediate L-arogenate (38) have been demon-strated in a number of microorganisms (10, 28),plants (4, 31), and fungi (1), and in Euglena

gracilis (6). One study revealed that tyrosinewas synthesized via arogenate in six diversespecies of cyanobacteria (33), and the evidencesuggested that all cyanobacteria might utilize thearogenate pathway as the exclusive, or primary,means of tyrosine biosynthesis. Differences inallosteric control patterns in the eubacteria havebeen reported in the biosynthetic pathways forthe aromatic amino acids (5, 22, 24, 35) and theamino acids of the aspartate family (8), and inthe oxidative dissimilation of aromatic com-pounds via the j-ketoadipate pathway (7).

Increasing evidence points to the conclusionthat such enzymological diversity in biosynthet-ic pathways may reflect ancient events of evolu-tionary divergence and that any particular pat-tern of enzymological steps to the end product(s)of a pathway is a conserved trait within ataxonomic group (5, 7, 8, 21, 22, 24, 25, 35).Recent enzymological patterning characteriza-tions in our laboratory with a large number ofpseudomonad species have led to the separationof five subgroups which parallel exactly the fivesubgroups defined by rRNA/DNA hybridization(5, 35). Although it has been generally conceded

65

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

66 HALL ET AL.

for years that only pragmatic microbial classifi-cations will ever be possible, a growing convic-tion exists that phylogeny-based microbial clas-sifications are possible after all. In addition tonucleic acid hybridization techniques and enzy-mological patterning, a third molecular tech-nique termed oligonucleotide cataloging (11) hasreceived considerable attention in recent years.To the extent that these independent techniqueshave overlapped in species coverage, agreementabout phylogenetic relationships has been im-pressively good.

It seemed likely, therefore, that a group aslarge and as ancient as the cyanobacteria wouldalso possess diversity in biochemical pathwaysor regulatory mechanisms, or both. To gain adeeper insight into the possibilities of enzymo-logical patterning within the cyanobacteria, weexamined the enzymology and allosteric inhibi-tion patterns of the aromatic amino acid path-way in 48 strains of cyanobacteria, includingrepresentatives from each of the five majorgroupings (sections) proposed by Rippka et al.(30).

MATERIALS AND METHODSOrganisms. Forty-five strains belonging to 13 form

genera and representative of the five major sections ofcyanobacteria proposed by Rippka et al. (30) weregrown at the American Type Culture Collection(ATCC) under the conditions recommended in TheAmerican Type Culture Collection Catalogue ofStrains I (14th ed., 1980), i.e., in Stanier medium BG11(30) (or medium MN [30] if a marine strain) at 26°C inshake flasks illuminated with fluorescent lights at anintensity of 2,000 Ix. Synechocystis sp. strain ATCC29108, Anabaena sp. strain ATCC 29151, and Syne-chococcus sp. strain ATCC 29404 (Agmenellum qua-druplicatum BG1) were grown at the State Universityof New York at Binghamton under conditions previ-ously described (15, 16, 19).

Cells were harvested at late log phase, pelleted, andthen frozen and shipped to New York packaged in dryice. Upon arrival, the frozen pellets were stored at-40°C before extract preparation. All strains exam-ined are listed in Table 3, together with PasteurCulture Collection (PCC) numbers. In a survey of thistype, it is not practical to optimize growth conditionsfor all the strains. In our experience, however, severegrowth conditions result in (i) cells with decreasedphycocyanin pigments and resultant yellow-green cellpellets, (ii) cells with a granular appearance that lyseeasily, (iii) both of the preceding effects, or (iv) nogrowth at all. The cell pellets from the ATCC weredark green, and microscopic examination of thawedcells revealed intact, healthy-looking cells and tri-chomes. Moreover, we have found that variable cul-ture conditions do not cause changes in specific activi-ties of enzymes of the aromatic pathway. Forexample, changes in the specific activity of 3-deoxy-D-arabinoheptulosonate 7-phosphate (DAHP) synthase,the only aromatic pathway enzyme so affected in thecyanobacteria to date, was limited to a threefolddifference, and that difference occurred only in mutant

strains, regardless of nutritional conditions that mightbe expected to cause changes in the specific activity(15, 16). Culture conditions, unless extreme, are un-likely to effect qualitative changes in the allostericbehavior or cofactor specificities of enzymes becausethese are intrinsic properties of the enzyme protein.

Extract preparation. Frozen cell pellets were thawedand suspended in 3 to 6 ml of cold 50 mM potassiumphosphate buffer (pH 7.5) containing 1 mM dithiothrei-tol. Cell suspensions were sonicated on ice at 70 mV inan Ultratip Labsonic system (Lab-Line Instruments,Inc.). The number of 30-s pulses required for celldisruption varied with the organism. Anabaena sp.strains required only three pulses, whereas strainswith well-developed sheaths required as many as eightpulses. Efficiency of cell disruption was determinedmicroscopically. After sonication, the suspensionswere centrifuged at 150,000 x g for 1 h at 5°C in amodel L5-65 ultracentrifuge (Beckman Instruments,Inc.). Small molecules were removed by passing theresulting supernatants through a Sephadex G-25 col-umn equilibrated with 50 mM potassium phosphatebuffer (pH 7.5) containing 1 mM dithiothreitol.

Analytical procedures. The assay conditions chosenwere those found to be optimal, or near optimal, forthe aromatic pathway enzymes and to produce resultsconsistent with in vivo patterns of regulation, asdetermined by mutant analyses, in Synechocystis sp.strain ATCC 29108 (15), Synechococcus sp. strainsATCC 27144 (unpublished data) and ATCC 29404 (20,33), and Anabaena sp. strain ATCC 29151 (16; G. C.Hall and R. A. Jensen, Curr. Microbiol., in press).When the strain being examined was capable of grow-ing under extreme environmental conditions, the assayprocedure was adjusted to reflect the particular condi-tion and is so noted in Results. Activities that were toolow to provide reliable data (results not included in thisreport) were attributed to low protein concentrations,i.e., less than 2.0 mg of protein per ml of extract,coupled with low specific activity. As a result of thehigh sensitivity of the dehydrogenase assays, howev-er, even extracts with low protein and low specificactivity provided sufficiently high readings to be reli-able. For example, at a final concentration of 250 Pg ofprotein per ml in the reaction mixtures, Synechocystissp. strain ATCC 29235 provided a fluorescence incre-ment/minute (AFU/min) reading of 1.58 ± 0.08 forarogenate/NADP dehydrogenase (specific activity,0.69).Crude extracts were used because of the nature of

this study and because partial purification results inthe inactivation of some enzymes, e.g., prephenatedehydratase of Synechocystis sp. strain ATCC 29108(unpublished data) and the tyrosine-sensitive DAHPsynthase isozyme ofAnabaena sp. strain ATCC 29151(16). Protein concentrations were determined by themethod of Bradford (3) as described in Bio-Rad Labo-ratories technical bulletin 1051. The method of Srini-vasan and Sprinson (32) as modified by Jensen andNester (23) was used to assay DAHP synthase. Reac-tion mixtures contained 50 mM potassium phosphatebuffer (pH 7.5), 0.5 mM dithiothreitol, 1 mM eachphosphoenolpyruvate and erythrose 4-phosphate, and100 to 700 pg of extract protein.Arogenate dehydrogenase, shikimate dehydroge-

nase, and prephenate dehydrogenase were assayed at37C by following the formation ofNADH or NADPH

J. BACTERIOL.

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

AROMATIC ACID BIOSYNTHESIS IN CYANOBACTERIA

spectrophotofluorometrically in an Aminco-Bowmanspectrophotofluorometer (excitation at 340 nm, emis-sion at 460 nm). Reaction mixtures contained 50 mMpotassium phosphate buffer (pH 7.5), 1 mM NAD orNADP, 1 mM shikimate or prephenate, or 0.5 mMarogenate, 0.13 mM dithiothreitol, and 50 to 100 pg ofextract protein. No interfering endogenous dehydro-genase activities were observed in any of the extracts.As an extra control extracts were also tested with acid-converted arogenate to eliminate any false activitiesdue to trace contaminants in the arogenate prepara-tion. The arogenate was treated with 0.1 N HCI for 10min at 37°C. This treatment results in complete con-version of arogenate to phenylalanine (38). The prepa-ration was neutralized with NaOH. No activity withacid-converted arogenate was found in any of theextracts. The tyrosine formed from arogenate viadehydrogenase activity was assayed fluorometrically(36) as previously described (15). The amount oftyrosine formed corresponded with the amount ofNADPH produced.Chorismate mutase and prephenate dehydratase

were determined by measuring the formation of phen-ylpyruvate at 37°C as previously described (28). Reac-tion mixtures contained 50 mM potassium phosphatebuffer (pH 7.5), 0.25 mM dithiothreitol, 1 mM choris-mate or prephenate, and 100 to 700 Fg of extractprotein. Arogenate dehydratase was measured as pre-viously described (28). Phenylalanine hydroxylase wasassayed as described by Guroff and Ito (14), with thefluorometric determination of tyrosine (36).

Biochemicals. Chorismate was isolated from culturesupernatants of Enterobacter aerogenes 62-1 as de-scribed by Gibson (12). Prephenate was isolated fromculture supernatants of Salmonella typhimurium (9).Arogenate (pretyrosine) was isolated from culturesupernatants of a multiple aromatic mutant (ATCC36373) of Neurospora crassa (26) and purified by amodified procedure (38). All other biochemicals wereobtained from Sigma Chemical Co.

RESULTSDAHP synthase. The first step unique to the

aromatic amino acid pathway is the condensa-tion of D-erythrose 4-phosphate and phospho-enolpyruvate to form DAHP as depicted in Fig.1. This enzyme activity was assayed and testedfor feedback inhibition by aromatic end productsand pathway intermediates. Cobalt ions werefound to stimulate activities in crude extracts byapproximately twofold and so were routinelyadded to reaction mixtures at a concentration of0.13 mM. Magnesium and manganous ions wereeither inhibitory or had no effect.The inhibition patterns for DAHP synthase

activities are presented in Tables 1 and 2. Thosestrains in which activities were too low to deter-mine meaningful regulatory patterns are notlisted in the tables. The predominant inhibitionpattern in the cyanobacteria was unimetaboliteinhibition by L-phenylalanine (retro-Phe), whichoccurred in over 55% of the strains examined.Only four examples of unimetabolite inhibition

by L-tyrosine (retro-Tyr) were found, and inhibi-tion in the presence of 0.1 mM tyrosine neverexceeded 65%.

Additive, cumulative, and concerted patternsof inhibition were found in several strains, allof which belong to the section IV groupingof Rippka et al. (30). Several of the form generain section IV (Table 2), including Anabaena,Nostoc, and Calothrix, contained strains withdifferent inhibition patterns. In the form genusSynechococcus (Table 1), both retro-Phe andretro-Tyr patterns were found. The implicationsof these differing inhibition patterns with respectto taxonomic divisions are presented below.Very little inhibition ofDAHP synthase activi-

ties by the pathway intermediates, phenylpyru-vate, prephenate, and chorismate, was ob-served. Exceptions included Anabaena sp.strain ATCC 29414, Oscillatoria sp. strainATCC 29134, Chroococcidiopsis sp. strainATCC 27900, and Synechocystis sp. strainATCC 27171 with 64, 60, 39, and 37% inhibition,respectively. In each of these, however, theDAHP synthase activity was highly sensitive toinhibition by phenylalanine, which ruled out thepossibility of a sequential mode of inhibition(24). No inhibition of DAHP synthase activityby the intermediate arogenate was observed.Dehydrogenase activities. Shikimate, prephen-

ate, and arogenate dehydrogenase activitieswere assayed for cofactor specificity as well asfor possible inhibitory effects by aromatic endproducts.

Shikimate and arogenate dehydrogenase ac-tivities are listed in Table 3. Arogenate dehydro-genase activity was found in all but three of thecyanobacteria surveyed. No arogenate dehydro-genase or prephenate dehydrogenase activitywas detected in Anabaena sp. strains ATCC27899, 29211, and 29414. Different assay condi-tions may be necessary for detection of activityin these strains. For all arogenate dehydroge-nase activities and the majority of the shikimatedehydrogenase activities, NADP was the pre-ferred cofactor. Many species exhibited a highdegree of specificity, and no shikimate dehydro-genase activity with NAD was detected inGloeothece sp. strain ATCC 27152 and Nostocsp. strain ATCC 29150. In contrast, NAD wasthe preferred cofactor for shikimate dehydro-genase activity in Synechocystis sp. strainsATCC 27171 and 27178 and Synechococcus sp.strains ATCC 29404 and 27180.

Significant prephenate dehydrogenase activitywas found in 16 different strains (Table 4). In 12of these strains, the activities had an absoluterequirement for NADP as cofactor, whereas inthe remaining four strains, the activity withNAD was less than 14% of the activity withNADP. Dehydrogenase activities with prephen-

VOL. 149, 1982 67

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

68 HALL ET AL.

ERYTHROSE 4-PHOSPHATE + PHOSPHOENOLPYRUVATE

3-DEOXY-D-ARABINOHEPTULOSONATE 7-PHOSPHATE

PREPHENIC ACID

DEHYDROSHIKIMIC ACID

SHIKIMIC TRYPTOPHAN

FIG. 1. Composite aromatic amino acid pathway including the various possible branchlets to phenylalanineand tyrosine that have been found among microorganisms. Enzymes: 1, DAHP synthase; 2, shikimatedehydrogenase; 3, chorismate mutase; 4, prephenate dehydratase; 5, prephenate dehydrogenase; 6, arogenatedehydrogenase; 7, arogenate dehydratase; 8, phenylalanine hydroxylase.

TABLE 1. Comparative allostery of DAHP synthase in sections I, II, III, and V cyanobacteria% Inhibition in presence of:

Genus PCC no. ATCC no. Sp act' Pheb Tyrb Trpb Aroc IntMdd

Section I

Section IIChroococcidiopsis 7203

Section IIIOscillatoriaLyngbya

Plectonema

Section VChlorogleopsis

Fischerella

2.90 8 (3)4.45 0 (0)0.76 0 (6)0.23 940.98 970.36 971.02 981.29 78

2.30 971.08 901.08 881.72 982.50 921.66 990.89 32

0.28 57

25 (73)0 (59)9 (0)177

251721

0 (0)0(1)0 (0)50000

20

0 40 210 23

100 095 1796 098 086 20

11 20 96 376 0 89 014 0 83 010 0 91 1511 0 93 2419 0 99 240 9 37 12

0 0 0 39

0.49 100 19 6 88 60

0.31 420.65 201.10 93

0 0 30 064 0 60 1211 0 87 0

2.74 92 15 11 87 17

0.61 1001.98 97

3.07 501.14 42

3 0 100 07 0 98 0

0 6 52 166 2 47 25

a Expressed as nanomoles of product formed per minute per milligam of extract protein.b Phenylalanine (Phe), tyrosine (Tyr), or tryptophan (Trp) present at a concentration of 0.1 mM. Values in

parentheses were obtained at effector concentrations of 0.5 mM.' Mixture of 0.1 mM each L-phenylalanine, L-tyrosine, and L-tryptophan.d Mixture of 0.5 mM each phenylpyruvate, prephenate, and chorismate.' Assayed at 50C. At 37°C the specific activity for Fischerella sp. strain ATCC 29161 was 0.37.

Synechococcus

Synechocystis

630169086911691067177202

7310975026702630867146902690570087509

271442714627192271912718029140294042917227171271502717829108291092911029235

27900

2913429119291212912627894

27193271812953829161e

7112640971147407

6306

69126718

75217414

J. BACTrERIOL.

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

VOL. 149, 1982 AROMATIC ACID BIOSYNTHESIS IN CYANOBACTERIA

U) U)

> > :>ig Y > > tN

V10~~~

u3

4 04E

42 gg E

u < <<<WMM WUK WSPz

00 0 ", I I I I,0I 000O

s W~IIIIW

_ O

1.& IIIIN 00 InN Un Nb

I--

NONHe,4,I en~-Cs~ 0,-; -

'n 00 N-

1. 00 %O rn 0 asN- -q %O 'f

e 1- -1o o-E- -

-

roNo

- - -r- t- W a,

" tn 0 o "t 0. = 0- en m -4

ON N WI,ej t_ _ _

I.-

oo m WI m as W) *fkn" mt-.q rI

m *nfi

0N %r eN 4

~0000 00E~

F- I

I I

lI I I

I II

0 'O NI%0oo o% et) N

en r t 0000 0 0

00

N-

VN IS %o eC,~ -- V-- N

00

1. C6

0

,'t Q = lqt Cfi oC

en en Io oo

A~ene- 0

W) lt C) " s " V mb xO o 14 o -

en 0^ m -o

0A C A 0 N 0

"~ 0%A A= V-V- n'Rien N1

N-

= (= C)

r Ftl e- 1% ir- t-

C,,

z- 1

.2 E*z

*P-X

69

.U

UE9r.,:g

i"'._

0 C.0 w.t>

2

0- - - c-

- --q -f

.=U

Q

U)*.a

C.)

U

J

0-

U)

:0

4-

aU,

EU

0

U,

0(A

0

itU

UU

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

TABLE 3. Comparison of shikimate and arogesfate dehydrogenase activitiesa with cofactors NAD andNADP

Shikimate ArogenateSection Genus PCC no. ATCC no. NAD NADP NAD NADP

I Synechocystis 6308 27150 0.78 6.40 0.58 2.26

Synechococcus

Gloeothece

II Chroococcidiopsis

III OscillatoriaPlectonema

Lyngbya

IV Anabaena

Nostoc

CylindrospermumScytonemaCalothrix

V Fischerella

Chlorogleopsis

0.690.364.201.630.210.060.341.660.150.241.472.640.102.230.45

0

0.26

1.171.070.250.731.100.47

1.180.521.320.320.160.040.030.05

00

0.240.670.16

0.270.181.140.180.340.220.15

0.240.14

0.120.04

2.47 02.15 03.73 0.541.28 0.180.48 0.060.20 0.08

3.10 0.731.26 0.220.82 0.201.50 0.358.25 3.1110.56 3.950.58 0.441.13 7.891.05 0.93

1.28 7.20

5.36 0.60

1.53 1.692.53 4.021.36 2.76

5.52 2.371.64 1.260.97 1.96

1.33 00.82 04.11 01.22 3.301.30 1.482.50 0.541.36 0.371.10 0.50

0 00.10 02.40 0.654.10 0.701.30 1.60

0.42 0.052.07 0.395.00 1.163.86 0.382.06 0.433.20 0.080.90 0.25

9.400.6910.003.463.843.582.873.403.582.2417.3322.251.76

11.702.05

20.20

8.52

4.6915.1018.006.495.6411.90

000

14.4016.528.747.946.699.621.726.655.9619.46

1.294.03

30.406.0010.286.583.30

1.42 0.39 3.660.34 0.24 1.16

0.61 0.11 1.141.26 0.25 1.17

a Activities expressed as nanomoles of product formed per minute per milligram of protein. A specific activityof "0" represents a flat line on the spectrophotofluorometer recording instrument (specific activity < 0.005).

b Anacystis nidulans strain Tx 20.c Agmenellum quadruplicatum BG-1.d At 50°C no shikimate dehydrogenase activity was detected, and the specific activity of arogenate/NADP

dehydrogenase was 1.28.

291082923527178271712910929110291722718027191271922714627144b2920329404C29140

27152

6902750967146702690570087502671769106911690863017009

731097202

6909

7203

711263066402740771146409

71226309

7118712064117119

742271076719

7310275247417

711071027103750763037504

75217414

69126718

27900

291342789427902291262912129119

27899292112941427892278932789829151294132913229150291052913329411

29204291712790127905291122915629158

2953829161d

2719327181

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

AROMATIC ACID BIOSYNTHESIS IN CYANOBACTERIA

TABLE 4. Dehydrogenase activities with prephenate as substrate

Genus PCC no. ATCC no. Prephenate/NADP As % of activit'dehydrogenase activitya with arogenate

Synechococcus 6301 27144 0.68 36908 27146 1.60 96911 27192 0.44 20

Synechocystis 6905 29109 0.04 16714 27178 0.09 17509 29235 0.33 48

Lyngbya 6409 29119 0.59 57114 29121 0.08 17407 29126 0.16 2

Anabaena 7120 27893 0.38 26411 27898 0.56 67119 29151 1.14 14

29413 0.26 4

Nostoc 6719 29105 1.37 2073102 29133 2.60 44

Fischerella 7512 29538 0.06 2a Activity expressed as nanomoles of product formed per minute per milligram of extract protein.b ArogenatelNADP dehydrogenase specific activities are listed in Table 3.

ate as substrate were generally very low whencompared with activities with arogenate as sub-strate. In Synechococcus sp. strain ATCC27192, Synechocystis sp. strain ATCC 29235,and Nostoc sp. strains ATCC 29105 and 29133,however, the prephenate/NADP dehydrogenaseactivities were at least 20%o or more of thearogenatefNADP activities. Whether the pre-phenate dehydrogenase activities were the resultof arogenate dehydrogenase substrate ambiguityor of the presence of a separate gene productcould not be determined from the available data.No consistent patterns of inhibition were ob-

served for any of the dehydrogenase activities.Several species possessed shikimate/NADP de-hydrogenase activity that was inhibited by amixture containing 0.5 mM each phenylalanine,tyrosine, and tryptophan (Table 5). In extractsof Gloeothece sp. strain ATCC 27152, Ana-baena sp. strain ATCC 27893, and Nostoc sp.strain ATCC 29150, the inhibition was greaterthan 50%o.

Inhibition of arogenate dehydrogenase by ty-rosine was infrequently observed (Table 6), andin only one case, Nostoc sp. strain ATCC 29133,was inhibition greater than 40%. In all species,inhibition occurred only when NAD was presentas cofactor. Because high concentrations (0.5mM) of L-tyrosine were used, a trivial decreasein activity would not be unexpected as a result ofmass action effects. No L-phenylalanine activa-tion of arogenate dehydrogenase or prephenatedehydrogenase was found.

Chorismate mutase. Chorismate mutase cata-lyzes the conversion of the branch-point com-pound, chorismate, to prephenate (Fig. 1). Verylittle regulation of chorismate mutase by L-phenylalanine and L-tyrosine, singly or in com-bination, was observed, even at concentrationsof 0.5 mM. Only in an extract of Cylindrosper-mum sp. strain ATCC 29204 was chorismatemutase significantly inhibited by L-tyrosine(94%o). L-Phenylalanine inhibition was observedin three strains: Gloeothece sp. strain ATCC27152 (100%), Anabaena sp. strain ATCC 27893(58%), and Calothrix sp. strain ATCC 27901(72%).

Chorismate mutase activities in two thermo-philic strains, Fischerella sp. strain ATCC 29538and Synechococcus sp. strain ATCC 27180,were assayed at both 37°C and 50°C. In Fischer-ella sp. strain ATCC 29538 extracts the activityof chorismate mutase at 37°C was 91% inhibitedby 0.5 mM L-tryptophan, whereas at 50°C theactivity was only 9% inhibited. In Synechococ-cus sp. strain ATCC 27180, at 37°C the activityof chorismate mutase was 100% inhibited by L-phenylalanine, but at 50°C it was only 10%inhibited.Prephenate dehydratase. Prephenate dehydra-

tase converts prephenate to the aromatic inter-mediate phenylpyruvate, which is subsequentlytransaminated to form L-phenylalanine (Fig. 1).In several extracts the presence of aminotrans-ferase activity interfered with the determinationof regulation by L-tyrosine and L-phenylalanine.

VOL. 149, 1982 71

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

TABLE 5. Inhibition of shikimate/NADP dehydrogenase activities by aromatic amino acidsAFU per min

Genus PCC no. ATCC no. Control Aroa % Inhibition

Gloeothece 6909 27152 0.54 0.10 81Synechocystis 6905 29109 1.50 1.00 33Plectonema 6402 27902 1.36 0.76 44Anabaena 7120 27893 15.00 6.26 58Nostoc 7107 29150 0.24 0.10 60Calothrix 7507 29112 3.60 2.58 29

a Aromatic amino acids, L-phenylalanine, L-tyrosine, and L-tryptophan, present at a concentration of 0.5 mMeach.

Deamination of L-tyrosine produces 4-hydroxy-phenylpyruvate which absorbs at 330 nm, anddeamination of L-phenylalanine produces phen-ylpyruvate. Attempts to specifically inhibit theaminotransferase activity were unsuccessful.However, corrections for the presence of 4-hydroxyphenylpyruvate could be made by read-ing the absorbance at wavelengths from 320 to330 nm. The X.. was indicative of the relativeamounts of phenylpyruvate and 4-hydroxyphen-ylpyruvate present in the assay mixture. Theformation of phenylpyruvate via phenylalaninedeamination often masked L-phenylalanine inhi-bition of prephenate dehydratase in the absenceof L-tyrosine, but the level produced was insuffi-cient to mask the inhibition of L-tyrosine activa-tion.

L-Tyrosine activated, and L-phenylalanine in-hibited, prephenate dehydratase activity in 56%of the strains examined. In many strains activityin the absence of L-tyrosine was below detect-able levels. In no instance was L-tyrosine activa-tion present without accompanying L-phenylala-nine inhibition. In Synechocystis sp. strainATCC 29108 inhibition by L-phenylalanine wascompetitive with respect to both activator (L-tyrosine) and substrate (unpublished data). Ineight strains, no prephenate dehydratase activitycould be detected even in the presence of L-tyrosine, or when an extract of one of these,Anabaena sp. strain ATCC 29151, was preparedin the presence of L-tyrosine. The absence of

activity was attributed to a high degree of en-zyme lability since, even in those extracts inwhich activity was present, the activity rapidlydiminished with time unless L-tyrosine was pres-ent. In three strains (Gloeothece sp. strainATCC 27152, Oscillatoria sp. strain ATCC29134, and Nostoc sp. strain ATCC 29411) pre-phenate dehydratase activity was observed, butL-tyrosine was ineffective as an activator and L-phenylalanine was an effective inhibitor. In Scy-tonema sp. strain ATCC 29171 the activity wasunchanged by the presence of either effector.

Arogenate dehydratase and phenylalanine hy-droxylase. Arogenate dehydratase catalyzes theconversion of arogenate to phenylalanine. Ofthe48 strains of cyanobacteria included in this sur-vey, 23 were assayed for arogenate dehydratase,but no activity was detected in any of thestrains. Similarly, no phenylalanine hydroxylaseactivity was found in any of these strains.

Implications of diversity. The cyanobacteria insections I and II are unicellular (single cells orcolonial aggregates), and sections III, IV, and Vcontain filamentous organisms. Grouping of thecyanobacteria into five sections by Rippka et al.(30) was determined solely by reproductive anddivisional characters, and further classificationinto generic groups was based on additionalmorphological characters. Although some cellu-lar characters may truly indicate taxonomic re-latedness, they are often hard to determine,being subject to variable cultural conditions and

TABLE 6. Inhibition of arogenate/NAD dehydrogenase activities by L-tyrosineAFU per min

Genus PCC no. ATCC no. Control Ty%t Inhibition

Synechococcus 6717 27180 1.92 1.56 186910 27191 2.10 1.27 406911 27192 4.22 3.20 23

Synechocystis 7509 29235b 1.58 1.03 35

Nostoc 73102 29133 1.78 0.42 77a L-Tyrosine present at 0.5 mM.b Cofactor was NADP. No activity was detected with NAD as cofactor.

72 HALL ET AL. J. BACTERIOL.

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

AROMATIC ACID BIOSYNTHESIS IN CYANOBACTERIA

interpretations. In the form genus Synechococ-cus, for example, are included 28 strains repre-senting seven different genera as defined byprevious classification schemes. This suggeststhat such a form genus could possibly be brokendown into subgenera by using more reliabletaxonomic characters such as nucleic acid se-quence homologies and enzymological pattern-ing, as has been done with the pseudomonads (5,35). Our data on aromatic amino acid biosynthe-sis served to delineate generic groupings in thecyanobacteria.The correlation of DAHP synthase regulation

with dehydrogenase data and DNA base compo-sition for strains of Synechococcus and Syne-chocystis is presented in Table 7. The moststriking feature was the separation that occurredwith regard to patterns ofDAHP synthase regu-lation.As a group, the various Synechocystis sp.

strains were characterized by potent inhibitionof DAHP synthase by L-phenylalanine (90% orbetter inhibition in the presence of 0.1 mM L-phenylalanine). In Synechocystis sp. strainATCC 29108 this sensitivity is accompanied bygrowth inhibition in the presence of exogenousphenylalanine (15). All the Synechocystis sp.strains exhibited an overwhelming preference ofarogenate dehydrogenase activity for the cofac-tor NADP. The low level of prephenate dehy-drogenase activity detected in strains ATCC27150 and 27178 could possibly be due to a smalldegree of arogenate dehydrogenase substrateambiguity. The low prephenate dehydrogenasespecific activities as compared with arogenatedehydrogenase activities (Table 4) support thisconclusion. Within the genus Synechocystis,Rippka et al. (30) recognized two subgroupswhich were distinguished by mean DNA basecompositions of low (35 to 37) and high (42 to 48)mol% guanine plus cytosine (G+C). Our exami-nation of enzyme activities in the aromatic path-way revealed the possible existence of a thirdsubgrouping listed as group 3 in Table 7. Thisgroup was distinguished by a complete lack ofarogenate dehydrogenase activity with the co-factor NAD. Synechocystis sp. strain ATCC29235, however, is distinguished from strainATCC 29108 by the presence of prephenatedehydrogenase activity and is further distin-guished from other Synechocystis sp. strains byits larger genome size (17) and its ability to growphotoheterotrophically at the expense of su-crose (30). In group 2 strains ATCC 27171 and27178 there was a slight preference for NADover NADP as the cofactor for shikimate dehy-drogenase. This preference, plus the differencein arogenate NADP/NAD activity ratios, wouldseem to preclude identity of strain ATCC 27171with strain ATCC 29109, as has been suggested

(17, 18, 30), but provides additional evidence forthe identity of strains ATCC 27171 and 27178.The strains designated as members of the

genus Synechococcus by Rippka et al. (30) fellinto two quite distinct groupings (Table 7).Group 1 was characterized by (i) a DAHP syn-thase mildly sensitive to tyrosine or not regulat-ed at all, (ii) significant prephenate dehydrogen-ase activities with NADP as cofactor, and (iii)very high mol% G+C. These strains are consid-ered by us to be representative of the genusSynechococcus.Synechococcus sp. strains in group 2 were

characterized by (i) sensitivity of DAHP synth-ase to inhibition by L-phenylalanine and (ii) theabsence of any prephenate dehydrogenase activ-ity in the presence ofNAD or NADP. Sequenceanalyses of 16S rRNA by Bonen et al. (2)indicate that Synechococcus sp. strains ATCC29404 and 29172 are less closely related toSynechococcus sp. strain ATCC 27144 than toSynechocystis sp. strain ATCC 27178, both ofwhich are considered reference strains for theirrespective genera by Rippka et al. (30). Indeed,according to the SAB values, a measure of relat-edness, Synechococcus sp. strain ATCC 29404is more closely related to Synechocystis sp.strain ATCC 27178 than is Synechocystis sp.strain ATCC 27179 (2). Synechococcus group 2was further subdivided into two groups, A andB, on the basis ofDNA composition as reportedby Herdman et al (18).The genera of section III include the non-

heterocystous, filamentous strains. Our surveyincluded only a few representative strains of thisgroup, but the enzymological data provided abasis for the subgrouping presented in Table 8.The Oscillatoria and Plectonema sp. strainswere characterized by retro-Phe inhibition ofDAHP synthase and the absence of any detect-able prephenate dehydrogenase activity. TheLyngbya sp. strains were characterized by thepresence of prephenate/NADP dehydrogenaseactivity, but, although two strains were charac-terized by the retro-Phe pattern of inhibition forDAHP synthase, one (ATCC 29121) possessed aretro-Tyr pattern of inhibition and most likelyrepresents an additional grouping within theLPP-B group as defined by Rippka et al. (30).Of the six generic groups in section IV (het-

erocystous, unisenate, filamentous forms), fourwere included in our survey. Inhibition patternsfor DAHP synthase were quite varied both with-in and between form genera. Table 9 correlatesthe DAHP synthase inhibition patterns withother enzymological characters and defines pro-posed groupings. Rippka et al. (30) separated theAnabaena sp. strains into two distinct groupsprimarily on the basis of cultural and morpholog-ical characteristics. Our enzymological data sep-

73VOL. 149, 1982

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

74 HALL ET AL.

-r~~~~ e~~~~ - e~~<

C, -4 r- VON -

NNn;44~!:N

I I e_e4r-t xo

0Z 0F "

0^e( 000 00CD

W0m u tn t m

CD xco 4n ON"f.6 .~ *~ *)*c-4 er-omm00H

*!~~~~ON QD0

1%0 £oC4 so OON ee e e4 F crON 0ro00 % OO e'I(6 %OO'l O6fi bi

X" "

F ,~.

000 0000 0z, 04 0000044Sggg boggoo

0N)rIt-

vph oo oo) N ,-4CDa ent- - V-4 -N-NVO%-4" -

No %O S%0o %

a" 4 4 a cos os

- -I

crcm F oa o%o%

XO- N0Oe

J. BACTERIOL.

I

04)C's

cEo

00

15>-

lo . E

C)

0c;

+0

0,

0202

C,,4-2:

0

0

C)

*g0

N

r

mb

z

U

Uq -

U .

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

AROMATIC ACID BIOSYNTHESIS IN CYANOBACTERIA

TABLE 8. Enzymological comparison of section III genera

Genus PCC no. ATCC no. DAHP synthase Prephenateinhibition dehydrogenasea

Oscillatoria 7112 29134 Retro-Phe 0

Plectonema 6306 27894 Retro-Phe 06402 27902 Retro-Phe 0

Lyngbya 7407 29126 Retro-Phe 26409 29119 Retro-Phe 5

Lyngbya 7114 29121 Retro-Tyr 1

a As percentage of activity with arogenate as substrate, and calculated from the data in Tables 3 and 4.

arated the Anabaena sp. strains into the same

groupings but with the possible addition of a

third grouping. Group 1 was characterized by an

additive pattern of DAHP synthase inhibition,the presence of prephenate dehydrogenase, anda relatively high specificity of shikimate dehy-drogenase for NADP as cofactor. Anabaena sp.

strain ATCC 29413 fit the above criteria exceptfor the apparent cumulative pattern of DAHPsynthase inhibition and thus may represent anadditional grouping (group 3). Strains assigned

to group 2 were characterized by the retro-Phepattern of DAHP synthase inhibition, the ab-sence of prephenate dehydrogenase activities,and increased relative activity of shikimate de-hydrogenase with NAD as cofactor. In three ofthe group 2 strains, no arogenate or prephenatedehydrogenase activity could be measured. Thismay be a result of some as yet undeterminedreaction requirement common to these strains.Anabaena sp. strains ATCC 27899 and 29211

were considered by Rippka et al. (30) to repre-

TABLE 9. Enzymological comparison of section IV generaDehydrogenase activities

Genus PCC no. ATCC no. DAHPsynthaseaGruinhibition Shikimatea Prepheteb Arogenatea GroupNADP/NAD phenate NADP/NAD

Anabaena 7120 27893 Additive 8.1 2 11.26411 27898 Additive 62.5 6 16.2 17119 29151 Additive 45.3 14 21.5

Anabaena 7118 27892 Retro-Phe 3.8 0 4.47122 27899 Retro-Phe 1.1 0 NAC6309 29211 Retro-Phe 1.6 0 NAc 2- 29414 Retro-Phe 3.1 0 NAC

Anabaena 29413 Cumulative 22.0 4 13.9 3

Nostoc 7524 29411 Additive 8.1 0 12.2 17422 29132 Additive NAC 0

Nostoc 6719 29105 Concerted 10.1 20 10.2 273102 29133 Cumulative 6.1 44 8.5 3

Calothrix 6303 29156 Retro-Phe 14.5 0 82.37504 29158 Retro-Phe 6.0 0 13.8 17102 27901 Retro-Phe 4.4 0 26.2

Calothrix 7507 29112 Cumulative 6.1 0 23.9 2

Calothrix 7103 27905 Concerted 21.4 0 15.8 3

Scytonema 7110 29171 Retro-Phe 11.5 0 10.3a Ratio of activity with NADP as cofactor to activity with NAD as cofactor, and calculated from the data in

Table 3.b As percentage of activity with arogenate as substrate, and calculated from the data in Tables 3 and 4.c No activity detected.

75VOL. 149, 1982

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

76 HALL ET AL.

sent the same species. DNA-DNA reassociationvalues also indicated a high degree of similaritybetween these strains (27). The retro-Phe pat-tern of DAHP synthase inhibition and the simi-larity of shikimatefNADP to shikimate/NAD de-hydrogenase ratios support this conclusion.However, strain ATCC 27892 was quite clearlysimilar to the strains in group 2 and was notidentical with the strains in group 1 as previouslysuggested (30). Strains ATCC 29151 and 27898appeared to be identical because of the similar-ities in dehydrogenase specific activities (Table3), DAHP synthase inhibition data (Table 2), andthe presence of prephenate dehydrogenase (Ta-ble 4). On the basis of these comparisons, how-ever, strain ATCC 27893 appeared to represent aspecies or strain different from strains ATCC27898 and 29151 because of the large differencein shikimate dehydrogenase NADP/NAD activi-ty ratios.Three groups were discerned in the form

genus Nostoc. Group 1 organisms were charac-terized by an additive DAHP synthase inhibitionpattern, group 2 by a concerted pattern, andgroup 3 by cumulative inhibition. It should benoted that in Nostoc sp. strain ATCC 29133 thearogenate/NAD dehydrogenase activity was in-hibited by L-tyrosine and there was a very highlevel of prephenate dehydrogenase activity.DNA-DNA reassociation values also show onlya moderate degree of interrelatedness betweenstrain ATCC 29133 and other Nostoc sp. strains(27). The absence of arogenate dehydrogenaseactivities with the cofactor NAD (Table 2) in twostrains in group 1 may characterize a fourthgrouping.

Strains in the form genus Calothrix separatedinto three groups on the basis ofDAHP synthaseregulation: group 1 characterized by the retro-Phe pattern, group 2 by the cumulative pattern,and group 3 by the concerted pattern. No pre-phenate dehydrogenase activity was found inany of the Calothrix strains.The base compositions of all strains in section

IV are in a very narrow range (38 to 44 mol%G+C) and provide no basis for subdivisions (18).

DISCUSSIONThe results presented in this report reveal that

the cyanobacteria possess a diverse array ofregulatory patterns for aromatic biosynthesis atthe first committed enzymatic step, accompa-nied by a more modest degree of diversity in theenzymology of subsequent reactions in the com-mon portion of the pathway (shikimate dehydro-genase) and in the diverging branchlets to tyro-sine and phenylalanine. It is perhaps notsurprising that the more sophisticated inhibitionpatterns for DAHP synthase (additive, cumula-

tive, and concerted) occur in those strains thatare considered to be more complex in that theyexhibit filamentous growth with the differentia-tion of heterocysts and akinetes. The additivepatterns ofDAHP synthase inhibition in severalof the section IV strains suggest the presence ofseparate regulatory isozymes. Indeed, mutantanalyses and physical separation of DAHPsynthase activities have confirmed the existenceof DAHP synthase regulatory isozymes in Ana-baena sp. strain ATCC 29151 (16).The diversity in subsequent reactions resides

primarily in the specificities of the shikimate andarogenate dehydrogenase enzymes for the co-factors NAD and NADP, and in the presence ofprephenate dehydrogenase activity in severalstrains. The NADP/NAD dehydrogenase activi-ty ratios should be useful in determining theidentity, or nonidentity, of separately isolatedstrains. It has not yet been determined whetherthe prephenate dehydrogenase activity is due toa separate enzyme or to a modest degree ofsubstrate ambiguity on the part of arogenatedehydrogenase. The results clearly indicate,however, that the arogenate pathway to tyrosineis a common conserved trait among the cyano-bacteria.

In an effort to correct and define cyanobacter-ial genera which had previously been classifiedwith the eucaryotic algae and so were character-ized on the basis of field observations and her-barium specimens, Rippka et al. (30) reclassifiedthe cyanobacteria on the basis of nutritional,physiological, and morphological characteris-tics. The cyanobacteria were grouped into fivemajor sections, and these sections were furthersubdivided into generic groups or "form gen-era." Many of these form genera very likelyconsist of more than one generic group, andbiochemical characterizations are needed to dif-ferentiate these groups. The enzymological pat-terns found in the aromatic pathway of thecyanobacteria, as reported here, did define sub-groups within the form genera Synechococcus,Synechocystis, Anabaena, Nostoc, and Caloth-rix. Consideration should be given to the possi-bility of grouping group 2 Synechococcus sp.strains with the enzymologically similar group 2Synechocystis sp. strains. These two groupsdiffer in cell shape and mode of cell division, butsuch characteristics are deemed to have littletaxonomic significance (11). Comparative analy-ses of 16S rRNA sequence homologies havebeen used to determine phylogenetic relation-ships among the procaryotes (11, 13), and Bonenet al. (2) have extended these studies to includeseveral strains of cyanobacteria. It was foundthat Synechococcus sp. strain ATCC 29404 andSynechocystis sp. strain ATCC 27178 are moreclosely related to each other than each is to

J. BACTERIOL.

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

AROMATIC ACID BIOSYNTHESIS IN CYANOBACTERIA

other strains in its own form genus. Consider-ation should also be given to reclassifying thestrains in section IV.

It should be noted that Synechococcus sp.strain ATCC 27144 (Anacystis nidulans) hasbeen the subject of many studies on cyanobac-terial physiology. Yet, according to its DAHPsynthase inhibition pattern and mol% G + C(18), this strain appears to represent a minoritygroup of cyanobacteria.On the basis of their studies, Bonen et al. (2)

postulated that filamentous forms of the generaNostoc and Fischerella arose from within theSynechocystis group. Our data revealed signifi-cant differences in DAHP synthase inhibitionpatterns between strains in the Nostoc andSynechocystis form genera. However, the enzy-mology of strains in group-2 Anabaena sp. isvery similar to that of the Synechocystis sp.strains, so it remains possible that these strainsarose from within the Synechocystis group. Thetwo Fischerella sp. strains examined did exhibitretro-Phe inhibition of DAHP synthase, which ischaracteristic of the Synechocystis sp. strains.Comparative 16S rRNA homology studies

have also indicated that the chloroplasts of thered alga Porphyridium cruentum are of definitecyanobacterial origin whereas Euglena chloro-plasts are possibly of non-cyanobacterial origin(2). Studies (6) have shown that in E. gracilisarogenate is an obligatory intermediate in thesynthesis of tyrosine, that arogenate dehydro-genase and DAHP synthase are highly sensitiveto inhibition by low levels of tyrosine, and thatarogenate can also serve as an intermediate inphenylalanine synthesis. The latter two charac-teristics differentiate Euglena from presentlyexisting cyanobacteria. In the study cited, how-ever, whole cell extracts rather than isolatedchloroplast extracts were used. In this context itis also interesting that mung bean seedlingscontain an unregulated arogenate dehydroge-nase activity in addition to a regulated prephen-ate dehydrogenase activity, and it was postulat-ed that the arogenate dehydrogenase might belocalized within the chloroplasts (31). Furtheranalyses of the aromatic amino acid pathway inred algae such as P. cruentum and plant chloro-plasts would contribute a great deal to ourunderstanding of the origins of chloroplasts andto the determination of the validity of the pro-posed polyphyletic origin of chloroplasts (29).

ACKNOWLEDGMENTSThis investigation was supported in part by Public Health

Service research grant AM-19447 from the National Institutesof Health and in part by funds from the Center for Somatic-cell Genetics and Biochemistry.We gratefully acknowledge the help of Carol Percik with the

illustrative art work.

LITERATURE CITED1. Bode, R., and D. BLrnbaum. 1979. Die Enzyme der

Biosynthese aromatischer Aminosauren bei Hansenulahenricii: Nachweiss und Charakterisierung der Enzymedes Pretyrosin-Weges. Z. Allg. Mikrobiol. 19:83-88.

2. Bonen, L., W. F. Doolittle, and G. E. Fox. 1979. Cyano-bacterial evolution: results of 16S ribosomal ribonucleicacid sequence analyses. Can. J. Biochem. 57:879-888.

3. Bradford, M. M. 1976. A rapid and sensitive method forthe quantitation of microgram quantities of protein utiliz-ing the principle of protein-dye binding. Anal. Biochem.72:248-254.

4. Byng, G. S., R. J. Whitaker, C. E. Flick, and R. A. Jensen.1981. Enzymology of L-tyrosine biosynthesis in corn (Zeamays). Phytochemistry 20:1289-1292.

5. Byng, G. S., R. J. Whitaker, R. L. Gherna, and R. A.Jensen. 1980. Variable enzymological patteming in tyrosine biosynthesis as a means of determining natural relat.edness in Pseudomonadaceae. 1. Bacteriol. 144:247-257.

6. Byng, G. S., R. J. Whitaker, C. L. Shapiro, and R. A.Jensen. 1981. The aromatic amino acid pathway branchesat L-arogenate in Euglena. Mol. Cell. Biol. 1:426-438.

7. Cinovas, J. L., L. N. Ornston, and R. Y. Stanler. 1967.Evolutionary significance of metabolic control systems.Science 156:1695-1699.

8. Cohen, G. N., R. Y. Stanier, and G. LeBras. 1969.Regulation of the biosynthesis of amino acids of theaspartate family in coliform bacteria and pseudomonads.J. Bacteriol. 99:791-801.

9. Dayan, J., and D. B. Sprlnson. 1970. Preparation ofprephenic acid. Methods Enzymol. 17A:559-561.

10. Fazel, A. M., J. R. Bowen, and R. A. Jensen. 1980.Arogenate (pretyrosine) is an obligatory intermediate ofL-tyrosine biosynthesis: confirmation in a microbial mu-tant. Proc. Natl. Acad. Sci. U.S.A. 77:1270-1273.

11. Fox, G. E., E. Stackebrandt, R. B. Hespell, J. Gibson, J.Manloff, T. A. Dyer, R. S. Wolfe, W. E. Balch, R. S.Tanner, L. J. Magrum, L. B. Zablen, R. Blakemore, R.Gupta, L. Bonen, B. J. Lewis, D. A. Stahl, K. R. Luehrsen,K. N. Chen, and C. R. Woese. 1980. The phylogeny ofprokaryotes. Science 209:457-463.

12. Gibson, F. 1964. Chorismic acid: purification and somechemical and physical studies. Biochem. J. 90:256-261.

13. Gibson, J., E. Stackebrandt, L. B. Zablen, R. Gupta, andC. R. Woese. 1979. A phylogenetic analysis of the purplephotosynthetic bacteria. Curr. Microbiol. 3:59-64.

14. Guroff, G., and T. Ito. 1965. Phenylalanine hydroxylationby Pseudomonas species (ATCC 11299a). J. Biol. Chem.240:1175-1184.

15. Hall, G. C., and R. A. Jensen. 1980. Enzymological basisfor growth inhibition by phenylalanine in the cyanobacter-ium Synechocystis sp. 29108. J. Bacteriol. 144:1034-1042.

16. Hall, G. C., and R. A. Jensen. 1981. Regulatory isozymesof 3-deoxy-D-arabinoheptulosonate 7-phosphate synthasein the cyanobacterium Anabaena sp. strain ATCC 29151.J. Bacteriol. 148:361-364.

17. Herdman, M., M. Janvier, R. Rfppka, and R. Y. Stanier.1979. Genome size of cyanobacteria. J. Gen. Microbiol.111:73-85.

18. Herdman, M., M. Janvler, J. B. Waterbury, R. Rippka, R.Y. Stanier, and M. Mandel. 1979. Deoxyribonucleic acidcomposition of cyanobacteria. J. Gen. Microbiol. 111:63-71.

19. Holtzclaw, W. D., L. F. Chapman, and C. S. Gowans.1972. Some regulatory properties of phospho-2-keto-3-deoxyheptonate aldolase from the blue-green alga Ana-cystis nidulans. Biochim. Biophys. Acta 268:562-572.

20. Ingram, L. O., and R. A. Jensen. 1973. Growth inhibitionby L-phenylalanine in Agmenellum quadruplicatum: aclue to some amino acid interrelationships. Arch. Mikro-biol. 91:221-233.

21. Jensen, R. A. 1976. Enzyme recruitment in evolution ofnew function. Annu. Rev. Microbiol. 30:409-425.

22. Jensen, R. A., D. S. Narr, and E. W. Nester. 1967.

VOL. 149, 1982 77

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from

78 HALL ET AL.

Comparative control of a branch-point enzyme in micro-organisms. J. Bacteriol. 94:1582-1593.

23. Jensen, R. A., and E. W. Nester. 1966. Regulatory en-zymes of aromatic amino acid biosynthesis in Bacillussubtilis. I. Purification and properties of 3-deoxy-D-ara-bino-heptulosonate 7-phosphate synthetase. J. Biol.Chem. 241:3365-3372.

24. Jensen, R. A., and J. L. Rebello. 1970. Comparativeallostery of microbial enzymes at metabolic branchpoints:evolutionary implications. Dev. Ind. Microbiol. 11:105-121.

25. Jensen, R. A., and S. L. Steamark. 1970. Comparativeallostery of 3-deoxy-D-arabino-heptulosonate 7-phos-phate synthetase as a molecular basis for classification. J.Bacteriol. 101:763-769.

26. Jensen, R. A., L. Zamir, M. St. Pierre, N. Patel, and D. L.Pierson. 1977. Isolation and preparation of pretyrosineaccumulated as a dead-end metabolite by Neurosporacrassa. J. Bacteriol. 132:896-903.

27. Lachance, M. A. 1981. Genetic relatedness of heterocys-tous cyanobacteria by deoxyribonucleic acid-deoxyribo-nucleic acid reassociation. Int. J. Syst. Bacteriol. 31:139-147.

28. Patel, N., D. L. Pierson, and R. A. Jensen. 1977. Dualenzymatic routes to L-tyrosine and L-phenylalanine viapretyrosine in Pseudomonas aeruginosa. J. Biol. Chem.252:5839-5846.

29. Raven, P. H. 1970. A multiple origin for plastids andmitochondria. Science 169:641-646.

30. Rippka, R., J. Deruelles, J. B. Waterbury, M. Herdman,and R. Y. Stanier. 1979. Generic assignments, strain

histories and properties of pure cultures of cyanobacteria.J. Gen. Microbiol. 111:1-61.

31. Rubin, J. L., and R. A. Jensen. 1979. The enzymology ofL-tyrosine biosynthesis in mung bean (Vigna radiata [L)Wilczek). Plant Physiol. 64:727-734.

32. Srlnvasan, P. R., and D. B. Sprison. 1959. a-Keto-3-deoxy-D-arabo-heptonic acid 7-phosphate synthetase. J.Biol. Chem. 234:716-722.

33. Stenmark, S. L., D. L. Pierson, R. A. Jensen, and G. I.Glover. 1974. Blue-green bacteria synthesize L-tyrosineby the pretyrosine pathway. Nuture (London) 247:290-292.

34. Weber, H. L., and A. BDck. 1968. Comparative studies onthe regulation of DAHP synthetase activity in blue-greenand green algae. Arch. Mikrobiol. 61:159-168.

35. Whltaker, R. J., G. S. Byng, R. L. Gberna, and R. A.Jensen. 1981. Comparative allostery of 3-deoxy-D-ara-bino-heptulosonate 7-phosphate synthetase as an indica-tor of taxonomic relatedness in pseudomonad genera. J.Bacteriol. 145:752-759.

36. Wong, P. W. K., M. E. O'Flyan, and T. Inouye. 1964.Micromethods for measuring phenylalanine and tyrosinein serum. Clin. Chem. 10:1098-1104.

37. Woololk, C. A., and E. R. Stadtman. 1964. Cumulativefeedback inhibition in the multiple end product regulationof glutamine synthetase activity in Escherichia coli. Bio-chem. Biophys. Res. Commun. 17:313-319.

38. Zamir, L. O., R. A. Jensen, B. H. Arlson, A. W. Douglas,G. Albers-Schnberg, and J. R. Bowen. 1980. Structure ofarogenate (pretyrosine), an amino acid intermediate ofaromatic biosynthesis. J. Am. Chem. Soc. 102:4499-4504.

J. BACTERIOL.

on June 4, 2018 by guesthttp://jb.asm

.org/D

ownloaded from