Biochemical genetics of hydrogen metabolism in Escherichia coli


Material Information

Biochemical genetics of hydrogen metabolism in Escherichia coli hydB, hydF, fhlA genes
Physical Description:
vi, 100 leaves : ill. ; 28 cm.
Sankar, Pushpam, 1956-
Publication Date:


Subjects / Keywords:
Escherichia coli   ( lcsh )
Hydrogen -- Metabolism   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1987.
Includes bibliographical references.
Statement of Responsibility:
by Pushpam Sankar.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001096542
notis - AFJ2232
oclc - 19484170
System ID:

Full Text

Escherichia coli: hydB, hydF and fhlA GENES







The author wishes to thank Dr. K. T. Shanmugam, his

major professor, for support, encouragement and for a

thorough training on how to think about scientific problems

and arrive at conclusions rationally.

The author also thanks Dr. D. E. Duggan, Dr. P. W. Chun,

Dr. J. F. Preston and Dr. W. B. Gurley for helpful advice

while serving on the author's advisory committee.

Special thanks to his wife, Srilatha, who joined him in

the last leg of this race, for patience.



ACKNOWLEDGEMENTS ................................ ii

ABSTRACT ............... ....................... v

INTRODUCTION ................................... 1

LITERATURE REVIEW ............................... 3

MATERIALS AND METHODS ........................... 11

Materials .................. ......... .. .... 11
Abbreviations ..... .... ...................... 11
Media and Growth Conditions ................ 14

Biochemical Experiments ................... 15

Whole Cell Enzyme Assays ................... 16
Enzyme Assays with Cell Extracts ........... 19

Genetic Experiments ....................... 20

Cloning the hyd Genes ....................... 20
Mapping the Genes in Recombinant Plasmids .. 22
Construction of Physical Map of Plasmid DNA
by Restriction Endonuclease .............. 22
Subcloning ................................. 28
Transposon Tn5 Mutagenesis of Various Genes
in a Recombinant Plasmid (pSE-128) ....... 31
Small Scale Plasmid Preparation ............ 32
Identification of Proteins produced from
Plasmid-Encoded Genes (Maxi-Cell Experiment) 33

RESULTS ....... ............................... 38

Cloning of hyd Genes ........................ 38
Subcloning of the Plasmid pSE-22 ........... 43
Characterization of hydB Gene ............ 46
-Identification of Gene-Product produced by
hydB Gene ............................... 60
Identification of a New hyd Gene, hydF ..... 64
Characterization of fhlA Gene .............. 73


DISCUSSION ..................................... 85

REFERENCES ................... .................. 93

BIOGRAPHICAL SKETCH ............................ 100

Abstract of Dissertation Presented to the Graduate
School of the University of Florida in Partial
Fulfilment of the Requirements for the
Degree of Doctor of Philosophy

Escherichia coli: hydB, hydF and fhlA "GENES


Pushpam Sankar

August 1987

Chairman: Dr. K. T. Shanmugam
Major Department: Microbiology and Cell Science

Mutant strains of Escherichia coli which lost

hydrogenase activity and hydrogen dependent functions were

found to carry mutations in two unlinked genetic loci,

namely, hydA and hydB. These mutations were mapped at 59 min

of E. coli chromosome. In this study, the hydB genetic locus

was studied in detail. The hydB locus contains four linked

genes, namely, hydB, hydF, fhlA and fdv. Gene-products of

hydB and hydF are essential for hydrogenase and hydrogen

dependent activities, while fhlA gene is needed for format

dehydrogenase-2 (FDH-2) and format hydrogenlyase (FHL)

activities. A cloned recombinant plasmid, pSE-128, carries

all four genes. The hydB gene is found to be contained

within a DNA segment of 800 basepairs and codes for a 32,000

dalton protein produced only under anaerobic conditions.


The hydF gene is about 1,000 basepairs in length. The

gene-product of hydF gene has not been identified. The

polarity of hydF gene has been proposed.

The gene-product of fhlA is needed for both format

dehydrogenase-2 (FDH-2) and FHL activities. This gene codes

for a 78,000 dalton protein which is produced under aerobic

and anaerobic conditions. The fdv gene codes for a 82,000

dalton protein whose function is not known. The fdv

gene-product is also produced both aerobically and

anaerobically. The direction in which fhlA and fdv genes are

transcribed has been determined. The physiological roles of

all these genes in H2 metabolism of E. coli are areas of

future research.


The ability to metabolize hydrogen gas is found in many

prokaryotes (3), and in some eukaryotes such as green algae

(71) and protozoa (42). The prokaryotic microorganisms

include several economically important bacteria, such as

methanogens, nitrogen-fixers, as well as fermentative

bacteria (3). Hydrogen is a metabolic by-product of many

anaerobic bacteria during fermentation of sugars (3).

Photosynthetic production of hydrogen by cyanobacteria and

green algae, using sun light and water, has been proposed as

a potential source of fuel and chemical feedstock (35).

Hydrogen also serves as a source of reducing power and

energy for both aerobic and anaerobic bacteria, under

conditions of energy limitation (3). Archaebacteria use

hydrogen as a reductant for the production of methane (5).

Reutilization of H2 by nitrogen-fixing organisms is also

known to enhance the energy efficiency of nitrogen fixation

process (18).

Inspite of the important role hydrogenase plays in

bacterial metabolism and its potential use in biotechnology

for fuel production and nitrogen fixation, very little is

known regarding the biochemical genetics of hydrogen

metabolism even in a well studied bacterium such as

Escherichia coli. Until 1985, there was no defined selection

procedure available for isolation of iqutant strains of

E. coli defective in H2 metabolism. This limited the genetic

studies on this important metabolic pathway.

The availability of a positive selection procedure for

isolation of mutant strains, defective in hydrogen

metabolism, made it possible to isolate a large number of

mutant strains altered in H2 metabolism in E. coli (40).

Analysis of these mutant strains helped identify several

genes involved in H2 metabolism in E. coli (54). These genes

are found grouped in four loci in the E. coli chromosome.

This study presents the biochemical and genetic

characterizations of the genes in one such locus.


In 1931, Stephenson and Stickland proposed the name

hydrogenase (EC 1.12) for the enzyme which catalyzes the

production and consumption of H2 (59). Since that time,

hydrogenase activity was demonstrated in several bacteria

(3), green algae (71), protozoa (42) and in higher plants,

such as barley (62). The protein had been purified from

several bacteria and characterized (3). The enzyme

hydrogenase catalyzes a reversible reaction as outlined


H2 + e- carrier (oxidized) == 2H + e- carrier (reduced)

Hydrogenase plays a vital role in the anaerobic

metabolism of microorganisms. In fermentative bacteria, in

the absence of exogenously added electron acceptors, H2

evolution by hydrogenase helps to oxidize the reduced

electron carriers produced during fermentation.

In E. coli, a facultative anaerobe, pyruvate metabolism,

yields format and acetyl CoA. The format is further

metabolized to produce H2 and CO2 by the enzyme complex

format hydrogenlyase which includes format

dehydrogenase-2, hydrogenase and intermediate electron

carriers (35). In Clostridium pasteurianum, a strict

anaerobe, pyruvate, generated from glucose catabolism, is

broken down to yield acetyl CoA, CO2 and H2 via

pyruvate:ferredoxin oxidoreductase and hydrogenase (3, 63).

In this organism, excess NADH generated at the

glyceraldehyde-3-phosphate dehydrogenase level is also

oxidized to produce H2 by NADH:ferredoxin oxidoreductase and


Hydrogenase also catalyzes the oxidation of H2 when H2

is the only source of reducing power and energy for the

cell. The electrons derived from the oxidation of H2 by

hydrogenase are shuttled through electron carriers to reduce

an ultimate electron acceptor such as fumarate in
E. coli (9) or SO4 as in the case of Desulfovibrio (46).
In Paracoccus denitrificans, nitrate can serve as the

terminal electron acceptor for the reducing power generated

from the oxidation of H2, while nitrate represses

hydrogenase activity in E. coli (58). Methanogens utilize H2

as a reductant to reduce CO2 to methane (5). This activity

of hydrogenase, generally called H2 uptake (HUP) activity,

generates energy for the growth" of the organism.

Photosynthetic bacteria also utilize H2 as a source of

reducing power to support CO2 fixation (32). Some of the

non-sulfur purple bacteria are known to evolve H2 in the

dark, fermentatively, as in E. coli (23). Both

photosynthetic bacteria and aerobic hydrogen-oxidizing

bacteria, such as Alcaligenes spp. can grow autotrophically,

with H2 as sole electron donor. Several members of the genus

Alcaligenes produce a cytoplasmic hydrogenase which can

reduce NAD directly. The NADH can be coupled to the

reduction of 02 to H20 or can serve as a reductant for CO2

fixation (57). Alcaligenes also produces a membrane-bound

hydrogenase, whose presumptive function in the cell is

02-dependent oxidation of H2 coupled to ATP production (19,

57). These few examples, indeed, demonstrate the importance

of hydrogenase(s) in the physiology and metabolism of a

diverse group of microorganisms.

The enzyme hydrogenase is usually associated with

electron carrier proteins and exists in the cell as part of

a multienzyme complex catalyzing unidirectional reactions.

However, hydrogenase, once purified, does catalyze the

oxidoreduction reaction reversibly, in the presence of

suitable, artificial, electron donors or acceptors (3, 34).

Hydrogenase from Clostridium pasteurianum was the first

to be purified to homogeneity (11). Since these initial

studies, hydrogenases from several other bacteria have been

purified (3). Based on the molecular weight and the number

of subunits, purified hydrogenases from several sources can

be grouped into three categories; i) a single polypeptide

chain with a molecular mass ranging from 50,000 to 66,000

daltons; ii) a protein comprising two non-identical subunits

with a native molecular mass ranging from 89,000 to 101,000

daltons; 62,000 to 67,000 daltons for a large subunit and

26,000 to 34,000 daltons for the small subunit; iii) a third

group of hydrogenase found in Paracoccus denitrificans and

Alcaligenes eutrophus is a tetramer with two large subunits

(63,000 to 67,000 daltons) and two small subunits (31,000 to

33,000 daltons).

Hydrogenase is a protein containing non-heme iron and

acid labile sulfur (3). Hydrogenases also contain nickel,

although exceptions to the presence of nickel had been

reported (3, 26). The presence of iron and nickel in the

holoenzyme dictates the need for processing of hydrogenase

apoprotein into holoenzyme. Hydrogenases, like other

non-heme iron and sulfur proteins, are usually sensitive to

oxygen and are irreversibly inactivated (3). Although the

hydrogenase protein from several organisms is capable of

tolerating 02, the enzyme from Clostridium pasteurianum and

Desulfovibrio gigas is extremely sensitive (11, 46).

Two hydrogenase activities have been demonstrated in

E. coli; H2 evolution (FHL hydrogenase) and H2 uptake (HUP

hydrogenase). Several investigators observed that crude

extracts obtained from E. coli cells, grown under different

growth conditions, exhibited multiple hydrogenase bands in

non-denaturing polyacrylamide gels stained for hydrogenase

activity, suggesting the possibility of iso-enzymes in the

cell (1, 69).

More recently, Ballantine and Boxer (6) using

immunological procedures, demonstrated two antigenically

distinct hydrogenase isoenzymes in E. coli containing

nickel. These investigators also proposed the presence of a

third hydrogenase isoenzyme in E. coli (55). Sawer and Boxer

(56) purified the hydrogenase isoenzyme-1 to near

homogeneity and found that it consisted of 2 polypeptides of

molecular mass 64,000 and 35,000 daltons. Ballentine and

Boxer (7) purified an active fragment of the second

hydrogenase and found that the native molecular mass of the

enzyme was 180,000 daltons, comprising 2 each of

non-identical subunits with molecular mass of 61,000 and

35,000 daltons. On the other hand, Adams and Hall (2), and

Patel (48) purified a hydrogenase, that is presumably the H2

uptake hydrogenase consisting of a single polypeptide with a

molecular mass of 56,000 daltons.

Studies on the genetics of hydrogenases and hydrogen

metabolism have only recently received great attention

(36, 50, 54, 70). In 1975, Pascal and her co-workers (47),

using a dye-overlay method, isolated and described a mutant

strain of E. coli which lacked hydrogenase activity and

mapped the mutation at 57 min of E. coli chromosome.

However, this mutant strain was found to be defective in

other enzymes in the metabolic pathway such as format

dehydrogenase-1 and -2. Graham et al. (24) isolated two

mutants of E. coli with similar characteristics, using a

similar dye-overlay procedure. Karube et al. (30, 31)

described a mutant strain of E. coli which also mapped at

this location. Krasna (36) isolated a hydrogenase mutant of

E. coli defective in all H2 dependent functions. A Mud l(Ap

lac) insertion mutation, termed ant (anaerobic electron

transport) was mapped at 59 min between mutS and srl genes

of E. coli. This strain lacked the ability to produce H2

(FHL), but had hydrogenase activity (70). Mutations in

pyruvate formatelyase (pfl) gene (13) or in a regulatory

gene, fnr (fumarate, nitrate reduction) (39), were also

found to influence the production of hydrogenase activity by

the cell.

Our laboratory developed and employed a unique positive

selection procedure to isolate a large number of mutants

defective in H2 metabolism in E. coli and separated them

into two major classes (40, 60). Class I mutant strains were

found defective in H2-uptake activity, but they possessed

format hydrogenlyase activity. The mutation in class I

mutant strains was mapped at 65 min of the E. coli

chromosome. The class II mutants lacked hydrogenase and all

the H2-dependent activities and the mutation was

mapped at 59 min. Subsequently, it was demonstrated by

cloning and complementation analysis that the class II

mutant strains can be further separated into two unlinked

groups, hydA and hydB. The hydB gene region was shown to

contain 3 genes, namely, hydB, fhlA and fdv (54).

Wu and Mandrand-Berthelot (68) isolated hyd mutant

strains and mapped the mutations at 77 min of E. coli

chromosome. Mutation in one of the genes, hydC, was

suppressed by exogenously added nickel while a mutation in a

nearby gene, hydD, was not. The function of the hydD gene is

unknown, at the present time. Waugh and Boxer (67) reported

a hyd mutation which abolished hydrogenase activity. This

mutation, which mapped in the hydB region, was also restored

to wild type character by the addition of nickel to the

growth medium. Later it was shown that a hyd mutant strain

(AK-23) isolated by Krasna (36) was similar to that of Waugh

and Boxer's strain (67) because the mutant phenotype of

strain AK-23 was also suppressed by exogenously added

nickel. This gene, involved in nickel metabolism, has been

termed hydE.

At least, seven hyd genes, whose gene products are

essential for hydrogenase activity, have been identified in

E. coli. Among them hydC and hydE are involved in nickel

transport into the cell. None of the other genes affecting

hydrogen metabolism, identified and studied in E. coli, was

shown to be the structural gene for hydrogenase, although

the structural gene for hydrogenase (H2-uptake hydrogenase)

has been isolated and characterized from other organisms

[Rhizobium japonicum (72), Desulfovibrio vulgaris

(Hildenborough) (65), D.gigas (12), and D. baculatus (44)].

The H2 metabolic pathway involves several proteins such

as pyruvate formatelyase, format dehydrogenases,

hydrogenases, fumarate reductase, ubiquinone, several

unknown electron carrier proteins and regulatory proteins.

Hydrogenases, format dehydrogenases, fumarate reductase and

some of the electron carriers are non-heme Fe and acid

labile sulfur proteins. Hydrogenase holoenzyme also contains

nickel and format dehydrogenase is a molybdo-selenoprotein.

Metals such as Fe, S, Mo, Se, and Ni have to be transported

into the cell and processed to convert the apo-proteins into

active holoenzymes. It is conceivable that these genes are

involved in the regulation of production and/or processing

of apo-hydrogenase to active hydrogenase in the cell. A

complete study of the genes involved in the regulation of H2

metabolism is an important step in understanding this




Biochemicals were purchased from Sigma Chemical Co.

Inorganic and organic chemicals were obtained from Fisher

Scientific Co. and were analytical grade. Restriction

endonucleases and T4 DNA ligase were obtained either from

Bethesda Research Laboratories (BRL) or New England BioLabs.

Calf intestinal alkaline phosphatase was purchased from

Boehringer Mannheim Biochemicals (BMB). Exonuclease Bal31

was purchased from either BMB or New England BioLabs.

Bacterial strains, bacteriophages and plasmids are

listed in Table 1. All bacterial strains are derivatives of

E. coli K-12.


HF medium (H2-fumarate) is a chemically defined medium

used to test the ability of E.coli cells to grow under

anaerobic conditions, utilizing H2 gas as electron donor and

fumarate as electron acceptor. HUP (H2 uptake) indicates the

ability of E. coli cells to use HF medium for growth. HUP

activity is defined as the utilization of H2 as an electron

Table 1.

Bacterial strains, bactriophages, and plasmids
used in this study

Strain Relevant genotype Reference
or phenotype



Puig 426

JC 10244



SE-5 3





MBM 7014

CSR 603

MC 4100


A 421

H~r PO2A relAl pit-10 tonA22 T2r

thi-l leu-6 suc-10 bioA2(?) galT27
rpsLl29 chlcT-

alaS cysC43 srl-300::Tnl0 thr-i
leu-6 thi-1 proA2 galK2 ara-14 xyl-5
mtl-1 lacYl his-4 argE3 rpsL31

JC 10244 alaS+

Puig 426 hydB103

Puig 426 hydAll0

Puig 426 hydF105

Puig 426 hydF107

JC 10244 hydB108 fhlAl01

SE-38 hydB+ fhlAl02::Tnl0

F araC(Am) araD (argF-lacU169)
trp(Am) mal(Am) rpsL relA thi supF

F thr-1 leuB6 proA2 phr-1 recAl
argE3 thi-1 uvrA6 aral4 lacYl galK2
xyl-5 mtl-i gyrA98TnalA9-87rpsL31
tsx33 "--supE44

araD139 A(argF-lac)205 flb85301
relAl rpsL150 deoCl A

b221 rex::Tn5 CI857 Oam23 Pam8

Table 1 continued.

Strain Relevant genotype Reference
or phenotype

Plasmid vectors

pBR-322 tet bla 10

pUC-19 bla lacZY 64

pSE-4 tet, bla, pSa origin of replication

donor with fumarate as the electron acceptor. Hup is denoted

to represent H2 uptake phenotype (Hup and Hup- represent

the wild type and mutant phenotypes, respectively) which

includes hydrogenase, fumarate reductase, and other electron

carriers required for H2 consumption. The gene symbol hyd is

used for hydrogenase, and hup is used to designate the genes

whose products are essential for H2 uptake. Gene symbol fdv

represents the gene product which is essential for format

dehydrogenase (FDH) activity that couples format oxidation

to reduction of artificial electron acceptor benzyl viologen

(BV). This format dehydrogenase (FDH-2) is a component of

the format hydrogenlyase (FHL) enzyme complex. The gene

symbol fhl is used to designate genes, the products of which

are needed for production of FHL activity.

Media and Growth Conditions

Luria broth was prepared as described previously (40).

Glucose minimal medium had the following composition:

Na 2HPO4, 6.25 g; KH2PO4, 0.75 g; NaCl, 2.0 g; FeSO4.7H20,

0.01 g; (NH4)2SO4, 1.0 g; MgSO4.7H20, 0.2 g; NaMoO4.2H20,

0.01 g; Na2SeO3, 0.263 mg in 1 liter of deionized water. The

final pH of the medium was 7.5. Glucose, autoclaved

separately, was present at a final concentration of 3.0 g/L

for aerobic cultures and 15.0 g/L for anaerobic cultures.

The HF medium was the same as described by Bernhard and

Gottschalk (9). Solid medium contained 15 g of agar per

liter of medium. Ampicillin, kanamycin and tetracycline,

when present, were added to the medium, after autoclaving,

to a final concentration of 100, 50, and 15 ug/ml,


For growth under H2, bacterial cultures were spread on

the surface of the HF medium in petri dishes, and the plates

were placed in a vacuum desiccator. The gas phase was

removed and replaced with H2, and the plates were incubated

at room temperature. Anaerobic conditions were established

in the desiccator within 18 to 24 hours, as determined by an

anaerobic indicator strip (Gas Pak; BBL Microbiology

Systems). Colonies were observed in about 5 to 7 days. All

cultures employed for enzyme assays were grown at 37C,

except the strains carrying alaS mutation. The alaS mutant

strains were grown at 30C.

Biochemical Experiments

Protein was determined by Folin phenol method, described

previously (40) with bovine serum albumin as the standard.

The enzyme activities are expressed as nanomoles of product

produced or substrate oxidized per minute per milligram of


Whole Cell Enzyme Assays.

Aerobically grown stationary phase cells were used as

inoculum [5% (vol/vol)]. The cells were grown anaerobically

in 20 ml of LB with 0.3% glucose in 16 X 150 mm tubes. The

tubes were filled to the top to achieve anaerobic conditions

and incubated, at 37C, for 4 h (temperature sensitive

mutant strains were grown at 30C, for 5 h). The cells were

collected by centrifugation, at 3,500 X g, for 10 min, at

room temperature and washed once with 5.0 ml of wash buffer

[phosphate buffer, 0.1 M, (pH 7.0) containing 1 mM reduced

glutathione and 100 ug/ml of chloramphenicol, to prevent

continued protein synthesis, during the assay]. The cells

were suspended in 1.0 ml of wash buffer and maintained in

ice, under N2 atmosphere. Hydrogenase and format

dehydrogenase activities were measured at 550nm as reduced

BV formation, using an UV-Vis spectrophotometer

(SLM-Aminco-DW2C) by monitoring the reduction of BV.

Hydrogenase. Hydrogenase activity was determined by two

different methods. First was an exchange reaction using 3H2

as an analog of H2. This reaction, 3H2 + H+ == 3H + 3H-H,

is independent of electron transport proteins (35) and

provides a direct measure of the hydrogenase activity

present in the cell. Hydrogenase activity was also

determined by the rate of reduction of BV by H2, in the

presence of hydrogenase, a common procedure used by other

investigators (47).

Tritium exchange assay. Whole cells, grown as described

above, were washed once with wash buffer and assayed at a

cell protein concentration of 50 to 100 ug/ml, in a tube (12

X 75 mm). The final assay volume was 0.2 ml. The tube was

sealed with a serum stopper and the gas phase was replaced

with helium. Into this tube, 0.7 ml of H2 gas was injected,

after removing 0.7 ml of helium. Tritium gas (25 ul; 11.2

mCi/mmol; New England Nuclear Corp.) was added to a final

concentration of 0.55 uCi per tube. After one hour of

incubation, at room temperature, the serum stopper was

removed, and the tritium gas was vented out in the hood for

10 min, after vigorous mixing of the tube contents.

Tritiated water present in a 50-ul sample was determined

with a scintillation counter, in Aquasol-2 scintillation

fluid. Hydrogenase activity was expressed as nanomoles of

tritiated water produced, per hour, per milligram of cell


Hydrogenase-dependent BV reduction. The reaction mixture

contained NaK-PO4 buffer (0.1 M; pH 7.0), BV (50 mM), and

the cell suspension in a final volume of 3.0 ml in an

anaerobic cuvette. The gas phase was replaced with H2. The

rate of BV reduction was determined at 550 nm.

Hydrogen uptake activities were also determined by

monitoring the disappearance of H2 from the gas phase by

using a gas chromotograph, in the presence of either BV or

fumarate, as electron acceptor. The assay mixture for these

reactions contained, in a final volume of 1.0 ml, NaK-PO4

buffer (0.1 M; pH 7.0), BV or fumarate (50 mM), and cell

suspension at a final concentration of 150 to 200 ug of cell

protein, in a 10 ml wheaton vial (40). The gas phase was

replaced with N2, and H2 was added to a final concentration

of 10%. The amount of H2 in the gas phase was determined at

different time intervals, with a Gas Chromatograph (Varian

model 920). The activity was expressed as nanomoles of H2

consumed, per minute, per milligram of cell protein.

FDH-2. The assay mixture contained, in a final volume of

3.0 ml, NaK-PO4 buffer (0.33 M, pH 7.0), BV (6.5 mM), sodium

format (0.1 M) and cell suspension. The rate of reduction

of BV was monitored.

FHL. The reaction mixture for FHL assay contained

NaK-PO4 buffer (0.33 M, pH 6.5), sodium format (0.1 M) and

cell suspension (250 to 300 ug of total cell protein), in a

final volume of 1.0 ml. The rate of H2 evolution with time

was monitored by using a gas chromatograph.

The enzyme activities were expressed as nanomoles of BV

reduced, per min, per mg of cell protein, using a molar

3 -1
extinction coefficient of 7.78 X 10 cm Formate

Hydrogenlyase activity and H2 uptake activity (HUP) were

expressed as nanomoles of H2 produced or consumed,

respectively, per minute per milligram cell protein.

Enzyme Assays with Cell Extracts.

Preparation of cells and extracts for enzyme assays.

Cells used for enzyme assays were grown in 1.0 liter of LB

medium, under anaerobic conditions, by filling the culture

vessel to the top. Aerobically grown stationary phase

LB-cultures were used as inoculum [5% (vol/vol)].

For preparation of extracts, cells from 4 h old cultures

were harvested by centrifugation at 8000 X g, for 10 min, at

4C. The cells were washed once with wash buffer and

centrifuged again at 12,000 X g (4C) for 10 min. The cells

were resuspended in 1.0 ml of wash buffer and passed through
a French pressure cell at 20,000 lb/in2. The broken cell

suspension was centrifuged at 20,000 X g, for 20 min, at 4 C

and the supernatant was collected. This crude extract was

maintained in ice, under an N2 atmosphere.

FDH-2. The assay mixture for FDH-2 contained NaK-PO4

buffer (0.33 M, pH 7.0), BV (6.5 mM), sodium format (40 mM)

and the cell extract. The final volume was adjusted to 3.0

ml with deionized water (21). The reaction was carried out

in an anaerobic cuvette, at room temperature, in an N2

atmosphere and the rate of BV reduction was monitored.

FHL. The reaction mixture for FHL assay contained

NaK-PO4 buffer (0.33 M, pH 6.5), sodium format (40 mM), and

cell extract (25) in a final volume of 1.0 ml. The reaction

was carried out at room temperature, with N2 in the gas

phase. The rate of production of H2 was monitored with a gas

chromatograph (Varian, model 920) equipped with a molecular

sieve 5A column.

Genetic Experiments

Cloning the hyd Gene. Total chromosomal DNA was isolated

from a prototrophic strain of E.coli K-12 strain K-10 (52).

Plasmid pBR322 (10) was used as a vector in the cloning

experiments. Plasmid DNA was isolated as described by Davis

et al. (16). A gene bank containing E.coli chromosomal DNA

fragments was constructed by the general procedures

described by Ditta et al. (17). Chromosomal DNA was

partially digested with the restriction endonuclease Sau3A

at a concentration of 0.1 unit of the enzyme per ug of DNA,

for 15 min, at 37C (54). The Sau3A digestion was stopped by

incubating the reaction mixture in a 65C water bath for 10

min. The digested DNA sample was layered on a 36 ml, 10 to

40% linear sucrose gradient. The sucrose gradient was made

with 9.0 ml each of 10, 20, 30, and 40% (wt/vol) sucrose

solution [containing Tris-HCl (20 mM, pH 8.0), EDTA (10 mM)

and NaCI (50 mM)] layered gently on top of each other

starting from 40% sucrose at the bottom of the centrifuge

tube. The sample was centrifuged in a SW 27 rotor, at 23,000

rpm for 18 h, at 25C. The gradient was fractionated by

collecting 1.0 ml samples, using an ISCO-density gradient

fractionator (Model 183). The DNA samples containing sizes

larger than 10 kb were pooled together. Agarose gel

electrophoresis was used to determine the size of the DNA

fragments, using HindIII-digested phage lambda DNA as

standard. The vector, plasmid pBR322, was digested with the

restriction endonuclease BamHl and the 5' end of the linear

DNA was dephosphorylated with calf intestinal alkaline

phosphatase (0.01 unit/pmol of 5'end of DNA for lh at 37C).

This dephosphorylated and linearized vector DNA was ligated

with Sau3A-digested chromosomal DNA of size above 10 kb,

using T4 DNA ligase. A hydrogenase-defective, recA strain of

E.coli, strain SE-61, was transformed (10) with the ligation

mixture, and the ampicillin-resistant colonies were selected

in LB medium supplemented with ampicillin. To isolate the

plasmids containing the hydrogenase genes, the ApR colonies

were transferred to HF medium by replica plating methods.
+ R
The Hup Ap colonies were selected and maintained. The

plasmid present in these clones was extracted and tested for

hyd and other relevant characters. Total plasmid DNA from

the pool of ApR transformants was also isolated separately

and maintained as an E. coli gene-bank. Starting with this

E. coli "gene-bank" DNA, plasmids capable of complementing

the hyd/hup mutants belonging to other classes and not

represented by strain SE-61 (e.g., strain SE-53) were also

isolated by the procedures described above.

Mapping the Genes in Recombinant Plasmids.

In order to identify the number of genes present in the

chromosomal DNA insert, physical maps of the plasmids, were

constructed based on restriction endonuclease digestion

pattern, using previously published procedures (54).

Restriction endonuclease digestion conditions were either as

described by Davis et al. (16) or as recommended by the

manufacturer of the enzyme.

Construction of Physical Map of Plasmid DNA by Restriction

Plasmid DNA can be identified by he presence of a

specific gene and also by the presence of unique restriction

endonuclease cleavage sites (33). Based on the location of

these unique sites in the linear DNA, different plasmid DNA

molecules can be differentiated. In this section, a strategy

used for constructing one such map is presented. Using this

general strategy, all other plasmids were characterized.

Fig. 1. Agarose gel electrophoresis of plasmid pSE-22
digested with various restriction endonucleases (see text
for details). The gel was cast with 0.8% agarose and
contained lug/ml of ethidium bromide. Lanes: 1, BamHI; 2,
EcoRI; 3, HindIII; 4, PstI; 5, SalI, 6, BamHI-HindIII; 7,
BamHI-EcoRI; 8, phage lambda DNA digested with HindIII; 9,
PstI-BamHI; 10, SalI-BamHI; 11, PstI-EcoRI; 12,
PstI-HindIII; 13, PstI-SalI; 14, SalI-HindIII and
15, SalI-EcoRI.

Plasmid DNA was digested with various restriction

endonucleases, independently and in combination with other

restriction endonucleases to generate DNA fragments. These

fragments were then subjected to electrophoresis in 0.8%

agarose gel. DNA fragments migrate in this gel based largely

on the size of the fragment. Plasmid DNA plasmidd pSE-22)

was digested with various restriction endonucleases as

described by Davis et al. (16). Procedures used for

preparation of agarose gel, and for electrophoresis were

described previously (43). In this example, an agarose gel,

showing the electrophoretic characteristics of DNA fragments

from restriction endonucleases-digested plasmid pSE-22, is

presented in Fig. 1.

Bacteriophage lambda DNA, digested by enzyme HindIII,

was used as a standard for molecular weight determination.

The enzyme HindIII digested the phage DNA at 7 locations and

released 8 fragments: 23.76, 9.46, 6.67, 4.26, 2.25, 1.96,

0.59, and 0.10 kilo base-pairs (kb). However, in the gel

only 6 fragments are visible. The last two fragments were

not detected because these fragments were smaller in size

for detection, under these experimental conditions. Using

the molecular weights of HindIII-digested phage lambda DNA

fragments and the relative migration (Table 2), a standard

curve was developed. Using this standard curve, the sizes of

restriction enzyme

Table 2. Enzyme HindIII digested phage lambda DNA (lane #8,
Fig. 1)

Distance Migrated

Fragment Size
(kilo base-pairs)













Table 3. Restriction fragments of plasmid pSE-22 generated
by restriction endonucleases (Fig. 1)

Enzyme(s) Distance Calculated Fragment
Lane# used migrated (cm) Size (kb)

1 BamHI

2 EcoRI

3 HindIII

4 PstI

5 SalI

6 BamHI-HindIII







7 BamHI-EcoRI 1.15
8 HindIII (phage lambda)

9 PstI-BamHI


11 PstI-EcoRI

12 PstI-HindIII

see Table 2.



















Table 3 continued.

Lane# Enzyme(s) Distance Calculated Fragment
used migrated (cm) Size (kb)

13 PstI-SalI

14 SalI-HindIII

15 SalI-EcoRI








fragments from plasmid pSE-22 (Fig.1) were deduced and are

listed in the Table 3.

In mapping plasmid DNA, the sizes of fragments generated

by a single restriction enzyme digest are compared with

sizes of fragments generated by two enzymes-digest. In the

two enzymes-digest, one of the enzymes is the same as used

in the single enzyme digest. In this above example, sizes of

DNA fragments generated by enzyme SalI (lane #5) was

compared with sizes of fragments obtained after

double-enzyme-digests, such as SalI-BamHI (lane #10);

SalI-EcoRI (lane #15); SalI-HindIII (lane #14) and SalI-PstI

(lane #13). From these results, the sites for other enzymes

(BamHI, EcoRI, HindIII, and PstI) in plasmid pSE-22 were

assigned in relation to SalI sites in the plasmid DNA.

Similarly, by comparing the sizes of fragments generated by

other enzymes, the sites of these restriction enzymes could

be obtained (Fig. 1, Table 3). By super-imposing these data,

a restriction map of plasmid pSE-22 was deduced (Fig. 3)i.

Subcloning. In all subcloning experiments, any one of

the vectors, such as, plasmid pBR-322 (10), pUC-19 (64), or

pSE-4 (this study) was used depending on the suitability of

a particular vector for a given experiment. DNA for ligation

was obtained by 95% ethanol precipitation after multiple

extractions with phenol:chloroform and ether. To the DNA

sample, half the volume of 7.5 M ammonium acetate was added,

mixed and to this 95% ethanol was added at 2 times the

aqueous volume. This sample was maintained at -200C for at

least 2 h before precipitating the DNA by centrifugation for

30 min, in an Eppendorf centrifuge, at 12,000 X g. After

draining the supernatant, 1.0 ml of 70% ethanol was added to

the tube to wash off the remaining salt. The DNA pellet

obtained after centrifugation was air dried. An appropriate

volume of sterile deionized water was added to dissolve the

pelleted DNA.

Subcloning was achieved primarily by two different ways.

The chromosomal DNA insert was partially digested by single

restriction endonuclease which has more than one recognition

site in the DNA insert and none in the vector part of the

plasmid, followed by self-ligation. After self-ligation, the

resulting plasmid would be deleted for different segments of

chromosomal DNA present in the plasmid (54).

Alternatively, the recombinant plasmid DNA was digested

with one or more restriction endonucleases and various

insert fragments were recloned. The DNA sample was subjected

to gel electrophoresis in a low-melting agarose (Sigma

Chemical Co.) gel. After electrophoresis, the desired band

of DNA was removed with the gel. The gel was heated to 65C,

for 10 min. To this sample, 2X volume of sterile deionized

water was added, mixed gently and extracted with equal

amounts of phenol saturated with Tris-HC1 (10mM, pH 8.0) and

EDTA (1mM) buffer, for 20 min. The sample was centrifuged,

at 12,000 X g, in an Eppendorf centrifuge, for 15 min, at

4 C. The top aqueous layer was extracted 2-3 times with

phenol:chloroform (1:1 vol/vol). The phenol was removed from

the DNA sample by extracting few times with ether. Ammonium

acetate (7.5 M) was added at a concentration of one half of

the volume. The DNA was precipitated with ethanol as above.

This DNA was ligated to appropriately digested vector


Physical mapping of genes by deletion was accomplished

by two different means. In the first procedure the

restriction endonuclease Sau3A was used to partially digest

the DNA fragments. The digested DNA was ligated with

appropriately digested vector DNA. Recombinant plasmids

obtained from E. coli cells transformed with this DNA were

characterized. In the second technique, exonuclease Bal31

was used to digest the insert DNA progressively from one

end, after linearizing the plasmid with an appropriate

restriction endonuclease (43). At various time intervals,

small samples of reaction mixture were removed, added to 200

ul of stop buffer EEDTA (0.1 M, pH 8.0), SDS (0.1%)] and

maintained for at least 5 min at 50 C to terminate the

reaction. Alternatively, the Bal31 digested DNA sample,

obtained as before, was subjected to gel electrophoresis in

low-melting agarose. The desired size fragments were removed

from the gel and purified as described before. The purified

DNA fragments were 'blunt-end' ligated (43), transformed

into appropriate mutant strains and the isolated plasmid DNA

was characterized.

Transposon Tn5 Mutagenesis of Various Genes in a Recombinant
Plasmid (pSE-128).

The recombinant plasmid pSE-128 was transformed into

E. coli K-12 strain MBM 7014 (supF). Phage lambda 421

carrying the transposon Tn5 was used to introduce Tn5 into

this strain by transduction (45). For these experiments, the

strain MBM 7014 was grown in 10.0 ml of LB supplemented with

maltose (0.3%), to a cell density of 2 X 10 CFU/ml. The

cells were centrifuged and resuspended in 1.0 ml of 10 mM

MgSO4. A 0.2 ml sample of this cell suspension was infected

with phage lambda 421 at a multiplicity of infection (MOI)

of 1.0, for 30 min, at room temperature. The cells were spun

down and washed once with 2.0 ml of LB. These cells were

suspended in 20.0 ml of LB with glucose (0.4%) and sodium

citrate (10 mM), in a 250 ml Erlenmeyer flask. The culture

was shaken at 200 rpm, for 30 min, at 37C. After this time,

30 ug/ml of kanamycin and 100 ug/ml of ampicillin were added

and the culture was shifted to 30 C and left shaking at 200

rpm, over night. Plasmid DNA was extracted from this culture

and used to transform hyd mutant strains. Transformants were

selected on LB-agar containing kanamycin (50 ug/ml) and

ampicillin (100 ug/ml). The KanR and ApR transformants were

tested for Hup and Fhl characteristics. Individual Hup and

Fhl colonies were selected. Plasmid present in these clones

were extracted and transformed into the same strain to

confirm the plasmid genotype.

Small Scale Plasmid Preparation. For routine small scale

plasmid preparation, two different procedures were used. The

first procedure is a modification of a method described by

Maniatis et al. (43). An overnight culture (1.5 ml) was

transferred to an Eppendorf tube and spun for 2 min. The

pellet was suspended in 0.1 ml of ice-cold Sucrose-EUTA-Tris

buffer (SET) with lysozyme. SET buffer contained Tris (25

mM, pH 8.0), EDTA (10 mM), and sucrose (50 mM). To the

required volume of this buffer, EDTA (1.0 M, pH 8.0) and

lysozyme were added to a final concentration of 0.1 M and 4

mg/ml respectively. The sample was mixed and incubated at

room temperature, for 5 min. After this period, 0.2 ml of

freshly prepared lysis buffer (0.2 N NaOH, 1% SDS) was added

and the tube contents were mixed gently for a few seconds.

After 5 min of incubation, in ice, 0.15 ml of potassium

acetate buffer was added, mixed gently and incubated for an

additional 5 min, in ice. The potassium acetate buffer

contained 60 ml of 5.0 M potassium acetate, 11.5 ml of

glacial acetic acid and 28.5 ml of water to the final

concentrations of 3.0 M potassium and 5.0 M acetate ions.

The pH of this buffer was adjusted to 4.8. The sample was

centrifuged in an Eppendorf centrifuge, at 12,000 X g, for 5

min, at 4C. To the supernatant, 0.05 ml of Ribonuclease A

(1 mg/ml, heat treated at 100C for 2 min to inactivate any

contaminating deoxyribonuclease) was added and incubated for

10 min at room temperature. This sample was extracted twice

with phenol:chloroform (1:1 vol/vol). The plasmid DNA was

precipitated with ethanol as before, after the removal of

phenol with ether.

In the second procedure, cells carrying plasmid DNA were

grown in 10 ml of LB over night (10). The cells were spun

down at 4C and suspended in 0.3 ml of Tris-HC1 (0.05 M, pH

8.0) buffer containing 25% glycerol. To this cell

suspension, 0.04 ml of 1.0 M EDTA and 1 mg of lysozyme were

added. The sample was incubated in ice for 10 min. About 1

mg/ml of pronase, 0.02 ml of 20% SDS, and 0.05 ml of

Ribonuclease A (1 mg/ml) were added to the spheroplast

preparation and incubated in ice for 10 min. The sample was

transferred to a small centrifuge tube for extraction with

an equal volume of cold phenol saturated in TE buffer (Tris

10 mM, pH 8.0; EDiTA 1 mM) for 10 min. The top aqueous layer

was obtained after centrifugation and extracted twice with

phenol:chloroform, ether and the DNA was precipitated with

ethanol as described before.

Identification of Proteins Produced from Plasmid-Encoded
Genes (Maxi-Cell Experiment).

The general procedure used for determining the proteins

produced by recombinant plasmids is as described by Sancar

et al. (53). E. coli maxi-cell K-12 strain CSR-603 was

transformed with the plasmids. The transformation was

achieved by a modification of the method described by

Bolivar et al. (10). The cells were grown in 10.0 ml of LB

at 30 C until the optical density corresponded to 5 X 107

CFU/ml. The culture was shifted to 37C until the cell

density reached 1 to 2 X 108 CFU/ml. The cells were spun

down at room temperature and washed once with 5.0 ml of 0.1

M NaCI. The cells were suspended in 5.0 ml of 0.1 M CaCl2

and incubated at room temperature for 20 min. The cells were

spun down again and resuspended in 1.0 ml of 0.1 M CaCl2 and

left in ice for 12 h. In a pre-chilled tube, 0.2 ml of cells

was taken and plasmid DNA was added at a final concentration

of 100 ng per tube. The tube contents were incubated in ice

for 30 min and shifted to 42C for 2 min. To this sample,

1.0 ml of LB was added and incubated at 37C for 1 h. The

cells were centrifuged and resuspended in 0.1 ml of LB and

plated on selective medium containing appropriate


The transformants (strain CSR 603 with plasmids) were

grown in K-9 medium (53) supplemented with MOPS buffer

(50 mM; pH 7.0). For aerobic expression of the genes in the

plasmid, 2.5 ml of aerobically grown culture

(2 X 108 CFU/ml) was transferred to a sterile petri-dish and

exposed to 1.4 uW/cm .sec of ultra violet irradiation (254

nm) for 50 sec. The intensity of UV radiation was measured

using an ultra-violet meter (DM-245N, Spectroline). The

irradiated culture was transferred to a 125 ml screw cap

flask. Freshly prepared cycloserine was added to the culture

at a final concentration of 200 ug/ml and shaken at 200 rpm

for 16 h.

For anaerobic expression of the genes in a plasmid, 10

ml of cells were grown in the same medium as above, but

under N2, in a 70.0 ml of serum-stoppered vial. The cells

were irradiated with UV (1.4 uW/cm .sec) under a constant

stream of N2 flowing over the culture. The irradiated

culture was transferred to a 25 ml Erlenmeyer flask

containing 20 ml of the same medium, pre-flushed with N2.

This medium also contained 200 ug/ml of freshly prepared

cycloserine. The flask was shaken (100 rpm) at 37C for 16

h. In the absence of shaking, cells settled down at the

bottom of the flask resulting in survival of

cycloserine-sensitive cells. After 16 h, the cells from both

cultures were centrifuged at room temperature and washed

twice with Hershey salts medium (53). The anaerobic culture

was handled with minimum exposure to 02. These cells were

suspended in 1.0 ml of Hershey medium and incubated for 1 h,

at 37C, to starve them for sulfur, in order to label the

cells with 35S-methionine (1,086 Ci/mmol, NEN). For

labelling with 35S-methionine, MgSO4 was replaced with MgCl2

and 50 ug/ml of required amino acids were added. The starved

cells were labelled with 5 uCi of 35S-methionine per tube

for 1 h, at 37C. The labelled cells were spun down and

washed twice with 1.0 ml of 0.1 M NaCI and resuspended in

0.1 ml of IX solubilization buffer (38). The 2X

solubilization buffer contained in 5 ml, 2.075 ml of Tris

(0.1 M, pH 7.0), 1.5 ml of 10% SDS, 1.0 ml of glycerol, 0.25

ml of beta-mercaptoethanol, and 0.005% bromocresol purple.

The lysed cells were incubated in a boiling water bath for 2

min and spun down in an Eppendorf centrifuge at 12,000 X g,

for 15 min. A 3 ul sample of the supernatant was removed and

counted in a scintillation counter after adding 2.5 ml of

scintillation fluid, ScintiVerse E (Fisher scientific Co.).

A sample of the extract with a radioactivity of 50,000

to 100,000 cpm was subjected to polyacylamide gel

electrophoresis (PAGE), in 12% acrylamide gels (38). After

electrophoresis, the gels were treated with either Enhance

or Enlighting (New England Nuclear) as outlined by NEN. In

some experiments, the gels were also treated with PPO using

the procedure developed by Jen and Thach (28). One side of

the gel was covered with Whatman 3mm chromatography paper

and the other side with all-purpose polyvinylchloride wrap.

These gels were vacuum dried at 80 C, for 1 h and 30 min,

using Eprotec senior (Haake Buckler) gel drier. The dried

gels were exposed to Kodak X-ray film, XR-5, in a Kodak

X-Omatic cassette with intensifying screens on both sides,

at -80 C, for few hours to 1 day as needed. The X-ray film

was developed using Kodak GBX-developer and fixer. Protein

molecular weight standards (Mark-VII, Sigma) were also run

in the same polyacrylamide gel, in a separate lane. The lane

with the molecular weight standards was cut out after the

run, processed and stained with coomassie blue as described

previously (38). The molecular mass of the radioactive

proteins were determined after comparing the Rf (relative

mobility is defined as the ratio between the distance of

protein migration and distance of tracking dye migration) of

the sample to the standards.


Cloning hyd Genes.

The class II mutant strains of E. coli, described

previously, lacked hydrogenase and were defective in all

hydrogen dependent activities (40). The mutation in these

strains was mapped between 58 and 59 min (srl and cysC) of

E. coli chromosome. It was demonstrated (40), by

bacteriophage Pl-mediated transduction, that these mutant

strains could be divided into two groups, namely, hydA and

hydB (Fig. 2). In order to further establish the fact that

these two operons are separate and independent, mutant

strains from representative groups were transformed with the

plasmid DNA from the "gene-bank" constructed and described

in the Methods section. The mutant strains, SE-53 (hydA101)

and strain SE-68 (hydBl08, fhlAl01) were plated on LB medium

containing ampicillin, after transformation with the plasmid

DNA from the "gene bank." Ampicillin resistant colonies were

transferred to HF medium by replica-plating techniques and

incubated for 5 days at room temperature, under H2

atmosphere. Individual HUP+ colonies were selected and

maintained. The presence of hyd/hup genes in the plasmid DNA

was confirmed by transformation of the same recipient strain

4 H
o H

.0d En


'hO a








4 r-4



by isolated plasmid DNA. These transformants were

characterized biochemically for enzyme activities, namely,

by the tritium exchange, by hydrogen dependent reduction of

BV and by fumarate-dependent H2-uptake. The results of these

experiments are presented in Table 4. The hydrogenase

production by the strain SE-53 was complemented by the

plasmid pSE-290 and not by plasmid pSE-111. The plasmid

pSE-111 was able to complement the strain SE-68 for

hydrogenase-dependent activities, while the plasmid pSE-290

failed to do so. The presence of the recombinant plasmids in

the wild-type parent, Puig 426, did not influence either the

hydrogenase or H2-uptake activity.

The physical map of the two plasmids (pSE-290 and

pSE-111), based on the restriction endonuclease digestion,

is presented in Fig. 3. This figure also contains the

restriction map of two other smaller plasmids, pSE-201 and

pSE-22, capable of complementing hydA and hydB mutant

strains, respectively. Plasmid pSE-22 carries a 9 kilo

base-pairs (kb) chromosomal DNA insert which is also present

within the 14 kb insert in the plasmid pSE-111. The plasmid

pSE-201 carries a 5.5 kb insert which overlaps part (4.8 kb)

of the 12.3 kb insert in plasmid pSE-290. Based on this

restriction analysis no overlap between the two groups of

plasmids plasmidss pSE-111 and pSE-201 or pSE-290) was

detected (Fig. 3). These results, obtained with the cloned








u0 o

C 44
0 0

S4) 0 0 0 -H
)>1 U $4

C 0 k 4JH 4-
-C O 0 O
0O 44 9 C U
O >i U 0-H 4 -H
U >
0M C 0"--,-4
04) -0
k "4 4" 0
0W 0 H )
U 44 "ri 0 U

0 4) -

0 A n E

() r-I C1 4
0,q10 04 a
e +> I $ "

m 0C 4 4) ( 0C (0
0 CO -4-I 0
4JC F!

U V r M -4 H

-4 r4 -4 ( ,Q $4
4 04Q $4 wc CE U

1- -

"r -




r4 N



) >











0 4
$ 4 0 I

'4 4J



0 0

l W


0 (d Z

) 0 r.I4 0

a N f

in ON Ln

m m
-4 -4
1- -4



DNA fragments, are in complete agreement with the genetic

evidence on the presence of two operons for hydrogenase

activity in the cell.

To further understand the organization of genes in the

hydB region, a fine structure analysis of the plasmid pSE-22

was carried out and presented below.

Subcloning of the plasmid pSE-22. Plasmid pSE-22 was

digested with restriction endonucleases SalI or PstI. The

results of this experiment are presented in Fig. 4. Complete

digestion of the plasmid pSE-22 with enzyme SalI produced

four fragments of sizes 1.15, 2.0, 2.8,and 7.8 kb. The

largest fragment contained the vector and a 3.45 kb fragment

of chromosomal DNA. Circularizing this DNA fragment with

T4-DNA ligase yielded plasmid pSE-127. The restriction

enzyme SAIl digestion of plasmid pSE-22, followed by

ligation produced plasmid pSE-128. Plasmid pSE-128 contained

the 2.8 and 7.8 kb fragments only. The 2.8 kb SalI fragment

present in plasmid pSE-128 was found to be reversed in

plasmid pSE-129. The orientation of this 2.8 kb fragment in

the plasmid was established after determining the location

of the KpnI and Clal sites in the fragment (Fig. 4). This

2.8 kb SalI fragment was also removed from plasmid pSE-22

and ligated into the SalI site of the vector plasmid pBR-322

(10). The resulting plasmids, pSE-125 and pSE-126 (Fig. 4)

differed in the orientation of the fragment and this was

(13.78 kb) pBR-322

pBR-322 P
Sal-1 / S
T4 -Ligose CK S

I I s

S- pSE-125 RS pSE-126 p pSE-130 p
(7.16kb) R (7.R6kb) R (8.93 kb)

B/So B/So

Fig.4. Isolation of plasmid derivatives from plasmid
pSE-127 (see Methods section for dpSE-129
(7.85kb) B/So (10.65 kb) (10.65 kb)

Fig.4. Isolation of plasmid derivatives from plasmid
pSE-22 (see Methods section for details)



u -4






0 )

U -H




M r-H r- in
-I r- in ^' 5 H
r-i 1rq in in ^ o r
HL1 H 1

Ho H
in 'o CA co ) c

rzI r-( rxf i- rzl r-

4 404 0404 04

*-" I
3 M





demonstrated by the location of Clal and KpnI sites in the

insert DNA with respect to the EcoRI and HindIII sites in

the vector. In constructing plasmid pSE-130, a 4.53 kb PstI

fragment which contained the 2.8 kb SalI fragment

internally, was ligated into the PstI site of the vector

plasmid pBR-322 (Fig. 4).

Characterization of hydB Gene

Strain SE-38 (hydB) was used in the study of hydB region

of DNA (Fig. 2) as a representative of several mutant

strains whose mutations were mapped in this region (54). The

plasmid derivatives obtained from plasmid pSE-22 were used

to transform strain SE-38 and the transformants were

selected on LB medium containing ampicillin or tetracycline

(for selecting plasmid pSE-130). The ApR or TetR

transformants were assayed for hydrogenase activity

(Table 5), using cell-free extracts of these transformants.

Plasmids pSE-125, pSE-126, pSE-128, pSE-129 and pSE-130

complemented the hydrogenase defect in the strain SE-38,

while plasmid pSE-127 failed to do so. These results suggest

that the 2.8 kb SalI fragment present in the plasmid pSE-125

(Fig. 4) contains the hydB gene. The same insert present in

plasmid pSE-126, although present in reverse orientation,

(Fig. 4) was able to complement the defect in strain SE-38,

pSE-125 I I I I

pSE-132 I I

Fig. 5. The chromosomal DNA insert present in plasmid
pSE-132. The KpnI-SalI fragment from pSE-125 was cloned
into plasmid vector pUC-19 (64). The vector in the plasmid
pSE-125 was plasmid pBR-322 (10). The vectors are not
shown. The restriction sites are C, Clal; K, KpnI; and
S, SalI. The numbers below the line represent the sizes of
DNA fragments in kilo basepairs.



%0 w
r-4 ro r-I
I ^" in r -
3 .* .

CO r' -I 0LAo
4 O

LA cwH '0 *

Oeqr -4 i 1-4 -n 5. *d


4- 0
r-I r-I 0 0 -,
(a p -,- +.40
4J M *04

1- 4 4 0 L 04

5i Q< 0
04 4 C eq
50 A CA (
A Ch co r-i 4
.in 0"N CNI r) I04
N 0) 0
0C 4 4 -H
0 rq
*CJ2 N 0 0) 0
004 .. r' ^ q U~

.1-Ieq eq o' 000
4-W 0 ** *04
woV Ln %0 q- LA g6
0 -'-4 N
04H 5 n i u 0 ri
0i m % r % (i Q

04 -4 0D$
r-Iv- 4 0 14

VQn in > C4
.9.4 eq CMC Y )O

-4 -4
0r40 0 I^ I E 3 .4.I

go0 000
,C)4 '-4 Ci) (I
r-- M, +>n
140 000.,-

0 .
'0 eq 000 o I>(

0C 01 0 00

'-4 H
.0 5.; -I I O C
(^o 4 :3 W

F0 rn S wdA
E-4 CO) 04 Ci) 10.QC

suggesting that the control region for hydB gene was present

within the 2.8 kb SalI fragment.

In order to "fine-structure" map the hydB gene various

methods, such as sub cloning and deletion analysis, were

sought. From the plasmid pSE-125, a 1.97 kb KpnI-SalI

fragment (Fig. 4) was removed and ligated into appropriately

digested vector plasmid pUC-19 (64). The resulting plasmid

pSE-132 (Fig. 5) was able to complement the hydB mutation

present in several strains and the biochemical

characteristics of strain SE-38 carrying plasmid pSE-132

are presented in Table 6. Complementation of the hyd defect

in strain SE-38 also restored both FHL and FDH-2 activities.

The enzyme activities presented in Table 6 were obtained

using whole cells.

In order to identify the location of the hydB gene,

within this 1.97 kb insert DNA fragment present in the

plasmid pSE-132 (Fig. 5), different regions of this insert

DNA were deleted. This was accomplished by partial digestion

of the 1.97 kb KpnI-SalI DNA by the enzyme Sau3A and

ligation of the resulting fragments to either KpnI and BamH1l

or SalI and BamHl-digested vector plasmid pUC-19. This

digestion-ligation experiment yielded several plasmids

carrying deletions from either side of the insert DNA. Some

of the plasmids constructed in this experiment are presented

in Fig. 6. Plasmid pSE-147 lost the smallest part (0.15 kb)







K B/Sa

K B/Sa

B/Sa S

B/Sa S

B/Sa s


Fig. 6. Chromosomal DNA inserts present in different
deletion clones obtained after the enzyme Sau3A partial
digestion of the KpnI-SalI fragment of plasmid pSE-132.
The vector plasmid used is plasmid pUC-19 (64) and is not
shown. Plasmid pUC-19 was digested with KpnI-BamHI for
construction of the plasmids pSE-143 and pSE-144 and
SalI-BamHI for constructing the plasmids pSE-147, pSE-148
and pSE-149 (see Methods section for details). The
restriction enzyme sites are B/Sa, BamHI-Sau3A junction;
K, KpnI; and S, SalI. The numbers below the lines
represent the sizes of DNA fragments in kilo basepairs.


r- Ln
r'4 C4 CV) ('4 41'
6 I (N a n C a' o- CV)
24 N in 0 n
10 33 . ,

a'l (4 Q) H i 5


0t Na


pa w
to z
so 04
H !B . gD

&< 4J 04 ^M as +
>i~ ~ n in o c



0) (4'4
N^ tm (Y) W Q4 Q Q 0
C ,g 4
0 3 (n Cn e r a o

4.rq4 a' C
0). a' CV) a

4JH r

34J ra C rV) o r-i co ON w > )
V e -4

I I I 4 44
r-44A 0 0 0 t
MfJ CV 4 04N 4 (4 m ra4 G
N 0 0. .
' S Q in '.0 in 4'< H N. (' 0 V
O -
H Hg

o Id IV0
V CV) "J r 0 3 3
0)0 -'. 3J '0 0~.9 .

4 4 4 44 4 I) e
00 04)4)


1d to I4 En to A >
'^ QO M M 0400(
4.1 --SQM CO C W

E' 1UI 02 01 .0 0 C

of the insert DNA. Plasmids pSE-143 and pSE-144 had retained

0.59 kb and 0.74 kb of the insert from the KpnI end,

respectively. The size of the inserts in the plasmids

pSE-148 and pSE-149 were 1.21 and 1.19 kb, respectively,

extending from the SalI end. Biochemical properties of the

strain SE-38 carrying these plasmids are presented in Table

7. The only plasmid that complemented the hyd defect in

strain SE-38 was plasmid pSE-147. All other plasmids failed

to complement the hydB mutation in strain SE-38. These

results suggest that a fragment larger than 1.21 kb is

necessary for this complementation.

Since the restriction endonuclease Sau3A has a unique

four nucleotide substrate specificity for digestion of the

duplex DNA, this method can only generate fixed unique

length deletions (33). On the other hand, enzyme Bal31, an

exonuclease, can progressively remove the nucleotides in the

linear DNA and this reaction can be controlled by the ratio

between the concentrations of DNA and the enzyme and also by

varying the time of incubation (43). Since partial digestion

of the DNA by enzyme Sau3A failed to yield the smallest

fragment of DNA needed to complement the hydB mutation, the

exonuclease Bal31 was used to generate deletions. However,

the transformants obtained from Bal31 deletion experiments,

starting with plasmid pSE-132 (vector plasmid pUC-19), grew

0 .45 0.38
pSE-131 |

I. 0 I/



1 07




I i




3 F
t I

h. I

Fig. 7. Chromosomal DNA inserts present in different
deletion plasmids obtained after exonuclease Bal31
digestion of plasmid pSE-131, linearized at the Clal site.
Plasmid pSE-162 was obtained by exonuclease Bal31
digestion of plasmid pSE-159, linearized at HpaI site. The
vector plasmid present in plasmids pSE-131, pSE-150,
pSE-151, pSE-152, and pSE-153 is plasmid pBR-322. The
vector plasmid present in plasmids pSE-159 and pSE-162 is
plasmid pSE-4. The part of vector plasmids are not shown.
The restriction sites are C, Clal; H, HpaI; K, KpnI; and
S,SalI. The numbers below the lines represent the sizes of
DNA fragments in kilo basepairs.












SC4 m Cl N .> Na -
iu * .
to P-
-oI 4 ,- 4 '-4 C'

r"O "O Om
X4.) -, 4- 4

0 0 004

N N LO n Ln L LA
'a *- E *


os o< E
:-0 A v LA N LA ON m C) V4

02 >4 ALA0

-I 0

V 2 M ...
44 l (0 LA LA : LA LA 4C
0 -4 r- -4 5 -4
4J H a, (

4- 1 '0a r- N C Y N $ > CNI

0 41 4 0
E0 04 04 0. 04 C)

poorly and the plasmids obtained from these transformants

had showed various DNA rearrangements. To circumvent these

problems, plasmid pSE-131 (Fig. 7) was constructed by

lighting the 2.35 kb ClaI-SalI fragment from plasmid pSE-125

(Fig. 5) into the vector plasmid pBR-322. Plasmid pSE-131

was used as the starting parent plasmid in this experiment.

The plasmid pSE-131 was digested with the enzyme Clal.

This linearized DNA was incubated with Bal31 for varying

lengths of time. The digested DNA was ligated, transformed

into strain SE-38 and ApR transformants were selected. In

four selected plasmids, pSE-150, pSE-151, pSE-152 and

pSE-153, the 2.35 kb DNA was found to be reduced to 1.50,

1.21, 1.12 and 0.98 kb, respectively (Fig. 7). Only plasmids

pSE-150 and pSE-151 restored the hydrogenase activity in

strain SE-38 (Table 8). Plasmid pSE-152 was smaller (about

0.1 kb) than the plasmid pSE-151 but failed to complement

the mutation in strain SE-38. Although plasmid pSE-148 (Fig.

6) carried a similar size DNA fragment (1.21 kb) as that of

plasmid pSE-151, it failed to complement the hydB mutation

(Table 7). These results suggest that the plasmid pSE-148

lacks the control region (promoter) for hydB gene

expression, whereas in the plasmid pSE-151 the control

region is present. Although plasmids pSE-147, pSE-150 and

pSE-151 complemented the hydB defect in strain SE-38 for

hydrogenase, they failed to restore FDH-2 activity to

parental level. This complementation analysis also indicates

that the plasmid pSE-148 is smaller than the plasmid

pSE-151, a small difference not detected by agarose gel

electrophores is.

In order to determine the other end of the hydB gene,

presumably the 3'end, the 2.8 kb SAlI fragment from plasmid

pSE-125 was used (Fig. 5). Attempts to delete from SalI end

had resulted in instability of the plasmid DNA and/or cell

lysis of the transformants containing these deleted

plasmids. The most probable cause of this result may be the

multi-copy nature of the plasmid vector, pBR-322. Since most

of the proteins that participate in hydrogen metabolism are

membrane-bound, production of a truncated protein, produced

from a plasmid, in which the gene is partly deleted, may

interfere with membrane architecture. In order to overcome

this problem, a low-copy plasmid vector was constructed

using plasmid vector pUCD-2 (14) as the starting plasmid.

The available low-copy vectors are much larger in size

and/or they have limited utility in this study, due to the

position and occurrence of restriction sites (51). Plasmid

pUCD-2 is 13 kb in size and carries the origins of

replication from plasmid colEl (14), a high copy-number E.

coli plasmid and pSA, a low copy-number plasmid, originally

isolated from Shigalla (66). This plasmid carries the genes

coding for kanamycin, ampicillin, spectinomycin and

tetracycline (14, 66). In the presence of pSa replication

origin this plasmid vector maintains itself at low-copies in

many gram negative bacteria (61). The genes conferring

resistance to kanamycin and spectinomycin and the gene(s)

responsible for colEl replication were deleted from the

plasmid pUCD-2 by employing various DNA endonucleases and

exonucleases. The resulting plasmid, pSE-4 is 4.45 kb in

size and contains only the pSa origin of replication. This

plasmid carries genes for ampicillin and tetracycline

resistance. The tet gene in this plasmid also has an

up-promoter mutation, so that transformants containing this

low-copy plasmid can be selected at higher tetracycline

concentration (greater than 5-7 ug/ml).

The 2.8 kb SalI fragment was ligated into the SalI site

of pSE-4 and the resulting plasmid was named pSE-159

(Fig. 7). The plasmid pSE-159 was linearized at the HpaI

site (Fig. 7) and enzyme Bal31 was used to delete the DNA

from this end. After ligation, strain SE-38 was transformed

with the DNA and both the HYD+ and HYD- transformants were

selected. Plasmid pSE-162 which was HYD was found to lack

about 0.5 kb (Fig. 7) and complemented the defect in the

strain SE-38 (Table 8). Among the exonuclease

"Bal31-deleted" plasmids which complemented strain SE-38 for

hydrogenase, only plasmid pSE-162 restored the normal FDH-2

activity, while other plasmids pSE-150 and pSE-151 failed to

0+1 ~

S N0 W E E-4
u4 u- U M 4-

in Os -- C En ,CC

r-4 -V4 )a -.4 -
S 0 l*'k4 i

E 4J 4 ,

U.O *- 0 m

E-4 04 *i Q* -I
0 0 4 0o *
0U m 4 O -4
O 4 # r4 >

4- .-u
OB Q g C-'-I

m g O-H

0 -4-i 4 *

4 Oi CO E 0

to 4 oa + 0

coQ 0 V

4o a1 V) f 0H -o
A4 f ) Aa
>1 4 > 4j
O r-i 40 0

*4 a a -,9 4 Z
0-40 o 4 0 -4

i1 (n 0o r- v O m 0 CN m O
rv L co r, r- A. A- o
. .
o q % r-4 4 J Co ,4
S- N r-q r-, -1
r-I (N H|

i-i r0-4 \

N No O0 "T'0 ~
i*o Mo C:) c> 00 -s4 s p
N4 (A ('; -4 ~
* .

CO i- vo ca Co -4o^'

* .

Co N s o o







:3 -40



4J 0

M 4J
0 d)


a ,



M r M r l r ^ f t)
N-I ,-I r-l ,-I ,-1

04 04 0404 04

04 wn
S,- I
0 Co

5N N

r- r-l





^J' m () N^ 0' (o m n %D rN

Ln ko o('4 C'4 r4 LA 0; N t '
N %o in rL N oo vAn (' N (A
NCo -4 COD
-4 '-4 -4 i-4



0 .
ft -u-

4)1- 04
4 04

S0 $4
go D 04

4 $4

14 >
ft V

4J M4 M0(

44 4.4 LWA
10 0) 0f


P-1 -4 -4 4J
0 0 0V
X 0 z0)
iU 3U'

3 0
+J C f






do so. The FDH-2 activity in these strains was twice the

amount present in the mutant strain alone.

In order to further establish that the hydB gene is

located at about 0.7 kb from the KpnI site (Fig. 7), plasmid

pSE-128 was mutagenized with transposon Tn5. The transposon

Tn5-mutagenized plasmid was used to transform strain SE-38

and the plasmids carrying Tn5 in the hydB gene were selected

(Table 9). All four plasmids failed to complement the hydB

mutation in strain SE-38 for hydrogenase and other

H2-dependent activities. The plasmid DNA from these clones

were analyzed to determine the position of transposon Tn5

(Fig. 8). Transposon Tn5 was located within the segment of

the DNA coding for hydB gene, as evidenced from enzyme Bal31

deletion studies (Fig. 7). These results suggest that one

end of the hydB gene lies at about 0.68 kb from the KpnI

site (Fig. 8). The other end of the hydB gene ends at about

0.5 kb from the SalI site (Fig. 7, Table 8).

Identification of Gene-Product Produced by the hydB Gene.

Using maxi-cell technique, it is possible to identify

the proteins produced by the genes encoded in a plasmid

(54). This technique is generally used to identify the

gene-protein relationship. For aerobic and anaerobic

expression of the genes in the plasmid, strain CSR-603

carrying plasmids were grown as described in the Methods

section. Plasmid vector pBR-322 produced two proteins as

identified by maxicell technique. The bla gene-product is a

28 kd (kilo dalton) ampicillinase protein and a 38 kd

protein is produced by the tet gene. In some experiments, a

31 kd protein was also detected among the proteins produced

by plasmid pBR-322, which is the precursor protein for

ampicillinase. These proteins are expressed both

anaerobically and aerobically (Figs 9, 10).

Plasmid pSE-125 produced two proteins; namely a 28 kd

ampicillinase protein and a 32 kd protein (Fig 9). The tet

gene product was absent because the chromosomal DNA

containing hydB gene (Fig. 4) was cloned into the SalI site

present in the tet gene. The proteins produced from various

subclones are presented in Fig. 9. The cells of strain

CSR-603, that carried these subclones were grown

anaerobically. Another protein with a molecular weight of 44

kd was detected in the extracts from the plasmid pSE-125

containing cells. Plasmid pSE-126 produced both the 32 and

44 kd proteins at higher levels than produced from plasmid

pSE-125, apart from the 28 kd ampicillinase protein. The 32

kd protein was not produced by CSR-603/pSE-127 since this

plasmid lacked the 2.8 kb SalI fragment carrying the hydB

gene (Fig. 4). Plasmid pSE-129, in which the 2.8 kb SalI

fragment was reversed as compared to plasmid pSE-128



45- 66-

36- 32 45-
29- 28

20-1- 36- 32
29- -2


Fig. 9. Protein products produced by plasmid-encoded genes.
Plasmids are subclones derived from plasmid pSE-22 (Fig. 4).
The cells harboring these plasmids were grown and processed
anaerobically (see Methods section for details). Molecular
weight of the proteins is in kilo daltons (kd). A, plasmid
pBR-322; B, plasmid pSE-125; C, plasmid pSE-126; D, plasmid
pSE-127; E, plasmid pSE-129; F, plasmid pBR-322 and
G, plasmid pSE-132.



--- 44

29 -- -29


, --24

Fig. 10. Protein products produced from plasmid-encoded
genes. The cells carrying these plasmids were grown
aerobically (see Methods section for details). Molecular
weight of proteins is in kd. A, plasmid pBR-322; B, pSE-125
and C, plasmid pSE-126.

(Fig. 4), also produced the 32 kd protein. Plasmid pSE-132

(Fig. 5), which carried only a 1.97 kb insert also produced

the 32 kd protein. When the cells harboring plasmid pSE-125

were grown aerobically the 32 kd protein was not produced,

suggesting that the hydB gene was expressed only

anaerobically (Fig. 10). These results suggest that the 32

kd protein is produced by hydB gene only under anaerobic

conditions. The origin of the 44 kd protein produced by

plasmids pSE-125 and pSE-126 is not known. The aerobically

grown cells of strain CSR-603/pSE-125 (Fig. 10) did not

produce the 44 kd protein, but a 22 kd protein appeared.

From the extracts of aerobically grown strain

CSR-603/pSE-126, 44 and 24 kd proteins were produced.

Identification of a New hyd Gene, hydF.

About 10% of the hydB mutant strains complemented by

plasmid pSE-125 (Fig. 4) were not complemented by plasmid

pSE-132 (Fig. 5, Table 10) for hydrogenase and hydrogen

dependent activities (data not shown). Strains SE-65 and

SE-67 were studied as representative of this new class of

mutants (Table 1). To further characterize the mutant

strains, plasmids pSE-131 (Fig. 7) and pSE-137 were used to

transform into these strains. Plasmid pSE-137 was

constructed by lighting the 0.83 kb SalI-KpnI fragment (not

present in plasmid pSE-132) from plasmid pSE-125 (Fig. 5)

into appropriately digested plasmid pUC-19. Biochemical











N-4 t

* 0


*. .-I




0 +

in "-4 C r in "-4 N C-
s m (n (q N mn CV) PI
M M M M M M M m
02 02 0 02 02 2 0 02
04040404 04040404


Ol %
04 02

M0 r- ( O V
* *
C>- Cl 00 flr> ifn "

C4 Ch

N ta r- n
"- 0 J' '0 0
* *

oi in

* *

0 i 0 0

Cl \o4 in 0 *

* *
ifl i^ 1% *- 1 "4 in
Cl i-l' r -l .
4 m"4 "4G

"41 Cl

qt ta CV) 0n co

H %0 N "4 r-

0 .^ 0

M Ln ko %0
"4 0. Cl .
N W m na

Q 0 Q Q Q
* *
"4 0 0 o


5 '.0 in '0 '.0
N '0 co in in






w 0


SM g 94
o *
$0 4 ~-


$4 0
$ 404
a a
0 O
a .5-I

4Jt $4


4J M 4
0 0 0


40 0 r.

0 0 0$

$4 C! l3




properties of two representative mutant strains with and

without these plasmids are presented in Table 10. The only

plasmid which complemented the hyd defect in strains SE-65

and SE-67 was the plasmid pSE-125. These results identify

the new gene and suggest that recognition site for KpnI lies

within the gene. Since other hyd genes, hydC, hydD (68) and

hydE (36) were already identified in E. coli, this new gene

is named hydF.

In order to further establish the existence of hydF

gene, segments from the 2.8 kb SalI fragment from plasmid

pSE-125 were deleted using enzyme Bal31. These deletion

plasmids exhibited DNA rearrangements and did not complement

the hydF mutation in either strain SE-65 or strain SE-67. In

order to eliminate the problems associated with plasmid

pBR-322 (a high copy-number vector plasmid), the low copy

vector plasmid pSE-4, described previously, was used.

Plasmid pSE-159 (Fig. 11), which carries the 2.8 kb SalI

fragment, was linearized at HpaI site and digested with

Bal31 for varying lengths of time. This Bal31-digested DNA

was ligated and used to transform the strain SE-38. The HYD

(as FHL-) transformants were selected. Plasmid DNA was

isolated from individual hydB- clones and transformed into

strain SE-65. The transformants were analyzed for both

hydrogenase and FHL activities. The smallest plasmid DNA

which complemented the hydF mutation in strain SE-65 was

pSE-159 I |
0.83 1.87

pSE-161 I I
S0.83 0.88

Fig. 11. Chromosomal DNA insert present in a deletion
plasmid derived from plasmid pSE-159 by exonuclease Bal31
digestion. Plasmid pSE-159 was linearized at the HpaI site
before digestion with enzyme Bal31. The vector is plasmid
pSE-4 and is not shown. The restriction sites are H, HpaI;
K, KpnI and S, SalI. The numbers below the line represent
the sizes of DNA fragments in kilo basepairs.


r- 5 r- CO o> r-
I H-1 in %0 c4

0 f- m co 0 ^r-
1z4 n(V r-4 r


O U r. a4
0 A P4 N0 N0- a)

Q ) IV 0
.0 *

$4 Oi J H$4
to .

:1 q HnmC
I *u
MS 0 r
W C 4 I-4
0OQ 04.e-40e
0 4 c4 M r W c 0

I44 44JQ
COH $, c oniQa40

0 4J0 4 C
0B 0 r4 OM)

'1i in 'o r- .c 04
0 S C 0 %0 % 404
t w r. 44 M 0.r4

+3 w ca ca 0 0 0
CM 04 0 4
0 r-I o
4u 4) 0 V 4

H- Nl H 0 0 a
5 a e 5 r< to in > <

(D S r4 in co to N r --
r-O LQ 01 %D %'s ri ^ 0
A $4 rq r$
M +i M M ,
00 r 0 C0 $


90 4J :3
H A4 4 w *o
040 04 3O CO O 000
(J040- 00L

H 0 OO '0 r' Q
$4 -H- i n < i 11 *
0 4-I .0 1 r-

UO 04 CO C)0

isolated and mapped with restriction endonucleases. Plasmid

pSE-161 carried an 1.64 kb DNA insert and had lost about

1.16 kb from the 2.8 kb SalI insert DNA present in the

starting plasmid pSE-159 (Fig. 11). Plasmid pSE-161

complemented the hydF mutation in strain SE-65 for

hydrogenase, HUP and FHL activities, while it failed to do

so in the hydB mutant strain, SE-38 (Table 11). These

results clearly show that the gene, hydF, is essential for

the production of hydrogenase in the cell and is adjacent to

the hydB gene.

In order to further demonstrate that hydF gene is

different from the hydB gene, because of the proximal

location of the two genes, plasmid pSE-128 was mutagenized

with transposon Tn5. The Tn5 mutagenized plasmid was used to

transform strain SE-65 and kan and ApR transformants were

selected. They were replicated onto HF medium and individual

HUP colonies were selected. Plasmids, defective in hydF

gene due to Tn5 insertion were isolated from these clones

and the location of Tn5 in the plasmids were mapped with

restriction endonucleases. Three of these plasmids were

chosen for biochemical characterization. In plasmids pSE-175

and pSE-176, the transposon Tn5 was mapped at about 0.17 kb

and 0.42 kb, respectively, from the left SalI site as

presented in Fig. 12. In plasmid pSE-177 the transposon was

located at about 0.51 kb to the right of the KpnI site as

presented in Fig. 12. Biochemical properties of strain SE-65

0 0 4J

M N.'.1

i 120 00*-

0C >

14O Q< 0 0
I 4 0 -H0
00 0 .

,-4 0
0 4 C

n C4 -,4

0 ",l N ,0 -," )
W 4 00 04 *0

o co -A 40

CO4 0 0 H

-n 4 W "i 0
(4 w M ci
0 0 to

OM 4 0 m
l- 4 4J ) 0
IVM *0 A
3 t; >1 4 S4 40
00 g a
LA 0 $ to $Q 0
co 0 $ 0 -H

4 4 44 4

o 00 0 Co
rn oM e *Mr40
0aN4 4 W 0 0
mo*1 0HM( ^






0 -









4W) co ^r .Mj oo
Sm w m
n wn v< N w


oo Co 'wj eq (

in N r- m
Z to c4 to
0\ C0 CD O0

I s (- o\ 1- (a
3 1 r-i cn LO r- m

r-4 N

V4 (- e o rto o4 P,
v~ N Co C4 1

rz qr oo in s in
in co ce r-4

NtCO0 M tst C4 Q
1 W) O oD C C? n

%0 H-4

wn in Mo OD to o
Nq Q N (1r) r Y

N~~ 0'0 to) CH0%
C% %0 H-4
r-4 r-

r-I NS
H e

in s
C Co

.1' U*
H eq

co Un '0 r- Co in '0 r.

04 040 4 04 ,4 04 04
I It It I I I I I

10 in
01 4

04 Co

04 r 4



0 0

04 (
ft .5-
4) 0

O O4

o v

4. 54 0)

54 >eN
u 0o

44- 4.4 44 A0
0 '0 0 g

4) 0 ) 0)

00 0

S Sk *
0.a ^ao

eq co eq
* *

m e
co to

C') Co o\
n (C e
kl C4 10
U3 in Q
Co r-
1* H H

carrying these plasmids are presented in the Table 12. While

plasmid pSE-177 failed to complement the hydF mutation in

strain SE-65, this strain with the other plasmids, pSE-176

and pSE-175 showed partial hydrogenase activity. The HUP and

FHL activities of these strains were proportional to the

levels of hydrogenase activity present in the cell. The

plasmid pSE-176 complemented the hydF mutation in strain

SE-65 for hydrogenase up to 24% of the activity as observed

with the same strain carrying the plasmid pSE-128 and the

HUP activity increased to 58 units from undetectable levels

(Table 12). Similarly, complementation by the plasmid

pSE-175 gave rise to 59% of hydrogenase and 20% of HUP

activities as compared with the strain SE-65 carrying the

plasmid pSE-128 (Table 12). The transposon insertion in the

hydF gene had also abolished FDH-2 activity in both strains

SE-65 and SE-67. Plasmids pSE-175, pSE-176 and pSE-177

complemented the hydB mutation in strain SE-38 for

hydrogenase and FHL activities. The FDH-2 activity was not

restored by these plasmids in strain SE-38 (Table 12).

The protein product produced by hydF gene was not

identifiable by maxi-cell technique. Plasmid pSE-125

(Fig. 4) contains the hydF gene (Table 10) but not plasmid

pSE-132 (Fig. 5, Table 10). The only proteins identified in

strain CSR-603 containing plasmids pSE-125 and pSE-132 were

the 28 Kd ampicillinase protein and the 32 kd hydB gene

product (Fig. 9). No other protein band which was present

with plasmid pSE-125 was missing in CSR-603/pSE-132. The

hydF gene-product may be produced in few copies and has

escaped the detection by maxi-cell technique.

Characterization of fhlA Gene.

Among the hydB mutant strains, one strain, SE-68

(hydBl08 fhlAl01, Table. 2) was found to be complemented for

hydrogenase activity, but not for FHL activity by plasmids

pSE-125 and pSE-130 (Fig. 4, Table 13). Plasmids pSE-128

(Fig. 4) and pSE-133 (Fig. 13), which carry a 3.7 kb Clal

DNA fragment from plasmid pSE-128 in the Clal site of the

vector plasmid pBR-322, were able to complement the mutation

in the strain SE-68 for both hydrogenase and FHL activities

(Table 13). Although plasmid pSE-130 (Fig. 4) had an

extension of 0.97 kb from the SalI end of plasmid pSE-125

(Fig. 4), it failed to complement the FHL defect in strain

SE-68. By extending the 0.97 kb SalI-PstI fragment present

in plasmid pSE-130 to about 0.38 kb (PstI-ClaI segment

present in plasmid pSE-133), normal FHL activity in strain

SE-68 was restored (Table. 13). Plasmid pSE-127, lacking the

2.8 kb SalI fragment (carrying the hydB gene) failed to

complement the defects in strain SE-68. These results

suggest that strain SE-68 carries two mutations; one in

hydB gene and another in a gene responsible for FHL activity.

This gene is named fhlA.

". "4

.r44 *
,4 )4H

o -H -4

w ft 4
41r4) M

44 J 4

4141H 4*1
0 14M
H-4 **4 0
< -H
zo x 4 4 0
4 H 0
*- .4

S0 4J 4 -4
S *- &

U U000 w
t o a rf A m*
4*J 4J
C4 4 u Ma

I | *r4 O ra 4 U I
c a ca

. * .
1 0 VO D 0 00

m r- C4 cq 0 v
9 4 9
to %0 C D H
r-4 < 0 r-f 1 C 4
n vo o
H' H eq








0 0



rO r- r S rM (

04 04 04 04


r4 I
3 W
04 CO

Q Q Q (N C
* .
0 0 0 '0

eq H

S Q -i ^ y
* *

H eq
H- H-

%D eq et
* .
's U.) '
m 00 cc

N C4
H- eq in
(r (u. H0
(* S S

In r! cO t)
N N cc (N









* ~~* H
(s) m~

o- t r-

n H

H Hn





M 4J
LS'4 0

04 J



M4 0 $4

4J14 04
14 > N
0 0 0)

t o r

44I r-I 44 .
a r 3
4) *
90 C)

000 N

*0 10 0 o

In order to further characterize this fhlA gene, it is

important to have a mutant strain carrying a mutation only

in fhlA gene, unlike strain SE-68 which is a double mutant.

A mutant strain, SE-1174 carrying a Transposon Tnl0

insertion mutation was constructed by a strategy outlined

below. Since the genes hydB and fhlA are adjacent to each

other, they should be easily co-transducible by

bacteriophage Pl. Strain SE-38 (hydB103) was used as the

recipient in this transduction experiment with bacteriophage

P1. A wild type E. coli strain MC4100 was randomly

mutagenized by Transposon Tnl0 and tetracycline resistant

colonies were selected. Bacteriophage Pl was grown in this

pool of TcR colonies. Strain SE-38 was transduced for TcR
R +
and the Tc transductants which were also hyd were selected

and screened for FHL activity. One such strain SE-1174, a

derivative of strain SE-38, was selected as hydB+ and FHL-.

Biochemical properties of this strain, SE-1174 with and

without the plasmids pSE-125, pSE-127, pSE-128 and pSE-133,

are presented in Table 13. Plasmids pSE-125 and pSE-127 did

not complement the defect in strain SE-1174 for FHL

activity, while plasmids pSE-128 and pSE-133 were able to

complement the defect in FHL activity. Attempts to clone the

fhlA gene by itself had resulted in instability and

rearrangement of the plasmids. Interestingly, a mutation in

the fhlA gene affected the FDH-2 activity and



SM E-4 M

I -4 0) 4 0
W 0 0
w-4 oH >o a i-I


E 1 .0 -4 4

Id 0 ) 4 J

n 4 **Jt.

EC 4 aW 4j a. -0

04 5 49 4J
.,.14 >0-- >
n 0NH 04.) 4 o

90 : 4) :1r.
4 Id0 14 0 0

0 0f3 C

$ 44 00 ) ,C-i

co 0 00
N 0 1 w 3 r4+
4 o H a a *-i >

0 Ol MC r4 *-4H

qJs Q Cl ~
. .
~ ~
"4 Cl

v %0
Q i Q N
. .
o ia 3 "4
"4 C-l


Ic- in
In tn

(4 "

0N m

"4 't9

ko m

* *4
OD a%
Co 0a

r-l "
N r-4
00 ^
u> n
* *




0 -,=
e c

0 V

0 2



0 o "4 Co m "-4
wl N co co
"4 "4 "4 4 "4 04

W 40 W W W W

N It
l r-
3 H
*9 IC
04 Co



0 tf
o *H


4 04

4 4)

AC r4

04 S

,C 4)
0 4
o V

0) 0U

14 0 0

4J W 0
9 4
14 > 0l
44 44 44

m (A m ,0

0 0 0 1

O 0 *

00 W

0 0 0 C

*~* 4 Cl *
"-4 LO P-
ta m
w H< cs










,v In 0A
to si 0
* *

tm P-4 M
'0n r>

1-4 "4

S in t-
C5 S

* *
Cl Cl "4-

complementation of fhlA mutation by plasmids pSE-128 and

pSE-133 (Table 13) had resulted in the restoration of FDH-2


In order to strengthen the identification of fhlA gene,

plasmid pSE-128 was mutagenized by transposon Tn5. The

mutagenized plasmid pool was used to transform strain SE-68.
R R +
Among the Ap and Kan transformants of strain SE-68, HYD ,

FHL clones were screened. From several plasmids which

carried Tn5 in fhlA gene, two were selected. The location of

Tn5 in these plasmids was mapped with restriction

endonucleases (Fig. 14). In plasmid pSE-180, the Tn5 was

located with in the 2.8 kb SalI fragment and this plasmid

failed to complement the defect in the strains SE-68 and

SE-1174 (Table 14). Plasmid pSE-181 which carries the Tn5

close to the Clal site (Fig. 14) restored FHL activity only

in strain SE-68 and not in SE-1174. These results suggest

that 2.8 kb SalI fragment contains part of fhlA gene with

part of this gene extending beyond the 0.97 kb SalI-PstI

fragment in plasmid pSE-130 (fig. 4). Plasmid pSE-133

contains the complete fhlA gene (Fig. 13, Table 13). Failure

to restore the normal FHL activity in strain SE-1174 by

plasmid pSE-181 may be due to the nature of the

transposition by Tnl0 in the fhlA gene. A truncated fhlA

gene product produced from the Tnl0 inserted chromosomal

fhlA gene could lead to the inhibition of the FHL activity

and such instances had been demonstrated previously (54).

When analysed by the maxi-cell technique, the plasmid

pSE-128 produced 32, 72 and 78 kd proteins apart from the 28

kd ampicillinase protein (Fig. 15). Plasmids pSE-125,

pSE-127 and pSE-129 did not produce the 78 kd protein and

did not complement the fhlA defect in the strains SE-68 and

SE-1174 (Table 13). Plasmid pSE-125 lacks part of the fhlA

gene as does plasmid pSE-127 (Fig. 4). In plasmid pSE-129,

the fhlA gene is split due to the reversal of the 2.8 kb

SalI fragment (Fig. 4). Plasmid pSE-133, which has the

complete fhlA gene (Fig. 13) produced a 78 kd protein (not

shown). These results show that fhlA gene produces a protein

of molecular weight of 78 kd. Plasmid pSE-130 (Fig. 4)

produced 66, 44, 38, 34 and 25 kd proteins (Fig. 16). The 38

kd protein is the gene-product of tet gene and the origin of

the 34 kd protein is not known. The 66 kd protein was

produced only in trace amounts as compared to the 44 and 25

kd proteins, which were produced in larger amounts. The 66

kd protein could be the truncated polypeptide from

incomplete fhlA, while the 44 and 25 kd proteins could be

the degradative products of the 66 kd protein. The 44 kd

protein appeared to be degraded further to a more stable 25

kd protein.

Plasmid pSE-111 produced 32, 78, and 82 kd proteins,

apart from the 28 kd bla gene product (Fig. 15). The 72 kd

protein produced from plasmid pSE-128 was absent among the

products produced from plasmid pSE-111. Plasmid pSE-111

^_ i


A 8


Fig. 15. Protein products produced aerobically by
plasmid-encoded genes. Molecular weight of the proteins is
in kd. A, plasmid pBR-322; B, plasmid pSE-125; C, plasmid
pSE-127, D, plasmid pSE-128, E, plasmid pSE-129 and
F, plasmid pSE-lll.








-- -44



Fig. 16. Protein products produced from plasmid pSE-130. The
strain CSR 603/pSE-130 was grown aerobically. The molecular
weight of the proteins is in kd.








Fig. 17. Protein products produced aerobically by
plasmid-encoded genes. Molecular weight of proteins is in
kd. A, plasmid pBR-322; B, plasmid pSE-128; C, plasmid
pSE-177 plasmidd pSE-128 h dF::Tn5), D, plasmid pSE-170
plasmidd pSE-128 hydB::Tn5) and E, plasmid pSE-180
plasmidd pSE-128 fhlA::Tn5).

carries a 2.1 kb extension from BamHI/Sau3A end of the

plasmid pSE-128 (Fig. 2). The 72 kd protein from plasmid

pSE-128 or the 82 kd protein from plasmid pSE-111 (Fig. 15)

are the products of the same gene fdv (54) and the plasmid

pSE-111 carries the complete fdv gene while plasmid pSE-128

does not. The gene-products of both the fhlA and fdv are

produced both.aerobically and anaerobically (Figs. 9, 15).

Figure 17 shows the protein products produced by plasmid

pSE-128 mutagenized with transposon Tn5. Proteins (62, 59

and 46 kd) not produced by plasmid pSE-128 were produced by

the transposable element Tn5 (14). The Tn5 insertion in the

hydB plasmidd pSE-170; lane D) or in the hydF gene plasmidd

pSE-177; lane C) did not affect the expression of fhlA or

fdv gene since the products of fhlA and fdv genes (78 and 72

kd, respectively) were produced from the plasmids pSE-170

and pSE-177. An insertion mutation in fhlA gene plasmidd

pSE-180; lane E) abolished the expression of this gene as

the 78 kd protein was not produced. These results show that

both the fhlA and fdv genes are independent operons.


Hydrogenase, the enzyme responsible for H2 metabolism in

E. coli, requires the presence of products from several

genes which are present at several locations in the E. coli

chromosome (59, 66, 77 min of the chromosome). Some of these

genes are involved in the production of hydrogenase

apoprotein, while others activate the apoprotein to

holoenzyme and/or regulate the synthesis of hydrogenase.

Since hydrogenase isoenzymes in E. coli are known to contain

iron and nickel (6), mutation in the genes which are

involved in the transport or processing of these minerals

will also affect the H2 metabolism. It was shown that when

E. coli cells were starved for iron, the format

dehydrogenase (FDH-2), hydrogenase and format hydrogenlyase

activities were lost (20) and addition of iron alone to these

starving cells allowed the production of all the three

enzymes. Since FDH-2, hydrogenase and electron carrier

proteins coupling these two enzymes are iron-sulfur proteins

(22), the above findings support the concept that

iron-deficient cells will not possess these enzyme


Mutations affecting nickel transport into the cells

abolished the hydrogenase activities (both FHL and HUP)

(68). Two genes are shown to be involved in nickel transport

into E. coli cells; hydC gene mapping at 77 min (68), and

hydE gene mapping at 59 min (36). The Hyd- phenotype in

these mutant strains of E. coli was reversed by the addition

of nickel to the medium.

The two format dehydrogenases present in E. coli are

molybdo-seleno proteins and thus need both molbdate and

selenite for activity (15). Formate dehydrogenase-1 (FDH-1)

was implicated in the nitrate respiration while format

dehydrogenase-2 (FDH-2) was shown to be part of format

hydrogenlyase system. Some of the chlorate-resistant strains

such as chlA and chlE (4, 27), identified as nitrate

respiration defective mutants, had low levels of format

dehydrogenase-2 and FHL activities (22). These mutants were

found to be defective in processing of molybdate (29).

Unpublished results from our laboratory showed that another

chlorate-resistant chlD mutant strain of E. coli, defective

in molybdate transport (27), produced undetectable levels of

FHL activity in the absence of high concentrations of Mo.

This phenotype was reversed by the addition of molybdate to

the medium. Recently, it was shown that a mutation in an

unlinked gene, fhlC, with genes involved in nitrate

respiration, abolished both the FDH-2 and FHL-hydrogenase

activities in Mo-deficient medium and the effect was

reversed by the addition of molybdate to the culture medium

(41). It is imperative that mutations affecting molybdate

metabolism will affect the FHL system since format

dehydrogenase-2 is part of that system. It is not known how

this regulates FHL-hydrogenase, although it is possible that

in the absence of FDH-2, the FHL-hydrogenase is not produced

or it is degraded once it is produced.

The pleiotropic regulatory gene, fnr, which maps at 29

min of the E. coli chromosome, was reported to control the

activities of nitrate reductase, fumarate reductase and

hydrogenase (39). A mutation in the fnr gene reduced the

hydrogenase activity to about 15% of the wild type levels

(39) and unpublished results from our laboratory, showed

that addition of molybdate could reverse the effect of fnr

gene-product on FHL production.

In this study, three genes are shown to be essential for

production of active hydrogenase in E. coli. These three

genes have been termed as hydA (54), hydB (54) and hydF.

Another gene, linked to the hydB gene (fhlA) is shown to be

required for production of FDH-2 and FHL activities. The

hydF, hydB and fhlA genes are located at 59 min of the E.

coli chromosome. The hydA and hydB genes are not contiguous.

The hydF and the hydB genes comprise adjacent operons.

Neither of the three hyd genes are found to be involved in

the transport of either nickel, a component of hydrogenase,

or molybdenum, a component of format dehydrogenase-2, since

addition of these compounds to the medium did not suppress

the effect of the mutation (data not shown). The fhlA gene

is linked to hydB gene.

A 2.8 kb SalI-fragment (Fig. 5) from E. coli chromosomal

DNA contains the hydB and hydF genes and part of the 5'end

of fhlA gene (Figs. 7, 11 and 14; Tables 8, 11 and 14).

Based on deletion analyses, it can be inferred that the 5'

end of the hydB gene lies at about 0.76 kb from the single

KpnI site (Fig. 7, Table 7) in the plasmid pSE-125, and the

3' end lies at about 0.5 kb from the SalI site in the

plasmid pSE-125 (Fig. 7). The hydB gene codes for a 32,000

dalton protein as observed by the maxi-cell experiment (Fig.

9) and is produced only anaerobically (Fig. 10).

The inability of the plasmid pSE-131 to complement the

strain SE-65 and SE-67 (Fig. 7, Table 9) for hydrogenase

activity identifies a new gene, hydF, so far unreported.

This gene is found to reside in a DNA fragment of 1.3 kb

(Fig. 8, 11 and 12). The increase in both hydrogenase and

HUP activities with the increasing size of hydF gene (not

interrupted by Tn5, Table 12) indicates that the 5' end of

the gene is near the KpnI site (Fig. 12) and the gene

proceeds through the KpnI site. The 3' end of the gene is

obviously closer to the left SalI site as presented in

Fig. 12.

A mutation in either hydB or hydF gene not only

abolished hydrogenase activity but also reduced the FDH-2

activity (15% of parent values) and complementation of

either of the mutations for hydrogenase, restored the FDH-2

activity (Tables 6 and 7). This suggests that a regulatory

pathway is coupling the FDH-2 and FHL-hydrogenase, since

both are components of FHL system. Interestingly, plasmids

carrying wild type hydB and hydF::Tn5 genes (Fig. 12),

although they complemented strain SE-38 (hydBl03) for

hydrogenase and FHL activities, the FDH-2 activities of the

strains carrying these plasmids stayed low plasmidd pSE-177;

Table 12). On the contrary, the plasmids wild type hydF gene

and hydB::Tn5 (Fig. 8) complemented strain SE-65 (hydFl01)

for hydrogenase, FDH-2 and FHL activities (Table 9). It was

shown previously that low levels of FDH-2 activity could

support normal FHL activity (54). These data suggest a

regulatory role for hydF gene-product on FDH-2 production

and activity. The hydF gene product is not identifiable by

maxi-cell technique, presumably, this protein is produced at

a low level in the cell. An expression vector system could

be profitably used to detect the gene-product produced from

the hydF gene.

A gene responsible for FDH-2 activity and therefore for

production of FHL activity in strains SE-68 and SE-1174 was

identified as fhlA by complementation analysis (Fig. 13,

Table 13). The deficiency in FHL activity may be a

consequence of Fdh2 phenotype. The plasmid pSE-130 (Fig. 4)

which failed to complement the defect in these strains

produced only a truncated gene-product of 66 kd (Fig. 16) as

compared to the full length 78 kd polypeptide. Extending the

partial fhlA gene present in the plasmid pSE-130 by another

0.4 kb led to the production of both the FDH-2 and FHL

activities (Table 13). This 0.4 kb DNA segment can easily

code for a 12 kd C-terminal part of the protein, in order to

make a complete 78 kd protein. These results indicate that

the direction of transcription of the fhlA gene is towards

SalI, PstI and Clal sites in the plasmid pSE-128 (Fig. 13).

Since the structural gene for FDH-2 was mapped at 92.4 min

of the E. coli chromosome (49), this fhlA gene can not be

the structural gene for FDH-2 enzyme. The nature of the

control the fhlA gene exerts on FDH-2 activity is unknown.

It was shown previously that the plasmid pSE-130

inhibited the FHL activity in the wild-type strain K-10

(54). Now it is clear that the truncated 66 kd protein could

lead to the inhibition of FHL activity in strain K-10. Since

the FHL system is a membrane complex, comprising FDH-2,

hydrogenase and electron carrier proteins, the truncated

protein may contribute to the disintegration of the complex

resulting in the loss of FHL activity. Another gene, termed

as fdv (54), is next to fhlA gene and produces a 82 kd

protein. Physiological role of this gene is not known.

Clearly, more data are needed to establish the roles of fhlA




00 -4


4 U)
ua 41

I -i-4
=3 .U

CN O3 (D
rn s to

* MSo
'-0 ole
44 0

4 4-J
jC .J
S N 41 *--4
C ad >
inE 4u *J

-- M U

Mi C




and fdv genes in regulating the FDH-2 activity and the FHL

system in E. coli.

Insertional inactivation of hydB and hydF genes by

transposon Tn5 did not affect the expression of fhlA or

fdv genes present in the plasmid pSE-128 (Fig. 17),

suggesting that these genes constitute independent operons.

A summary of the results has been presented in the Fig. 18.

It is only known that the hydB and hydF gene-products are

essential for hydrogenase and all H2 dependent activities.

Their regulatory roles are not understood. Both biochemical

and genetic characterizations of the strains carrying lacZ

fusions in each of the three genes at chromosomal level, may

yield some information on the nature of the control these

genes exert on hydrogenase.


1. Ackrell, B. A. C., R. N. Asato, and H. F. Mower. 1966.
Multiple forms of bacterial hydrogenases. J. Bacteriol.

2. Adams, M. W. W., and D. 0. Hall. 1979. Purification of
the membrane-bound hydrogenase of Escherichia coli.
Biochem. J. 183:11-22.

3. Adams, M. W. W., L. E. Mortenson, and J.S. Chen. 1981.
Hydrogenase. Biochim. Biophys. Acta. 594:105-176.

4. Bachmann, B. J. 1983. Linkage map of Escherichia coli
K-12, edition 7. Microbiol. Rev. 47:180-230.

5. Balch, W. E., G. E. Fox, L. J. Magrum, C. R. Woese, and
R. S. Wolfe. 1979. Methanogens: reevaluation of a unique
biological group. Microbiol. Rev. 43:260-296.

6. Ballantine, S. P., and D. H. Boxer. 1985. Nickel
containing hydrogenase isoenzymes from anaerobically
grown Escherichia coli K-12. J. Bacteriol. 163:454-459.

7. Ballantine, S. P., and D. H. Boxer. 1986. Isolation and
characterization of a soluble active fragment of
hydrogenase isoenzyme 2 from the membranes of
anaerobically grown Escherichia coli. Eur. J. Biochem.

8. Berman, M. L., and J. Beckwith. 1979. Fusions of the lac
operon to the transfer RNA gene tyrT of Escherichia
coli. J. Mol. Biol. 130:285-301.

9. Bernhard, T., and G. Gottschalk. 1978. Cell yields of
Escherichia coli during anaerobic growth on fumarate and
molecular hydrogen. Arch. Microbiol. 116:235-238.

10. Bolivar, F., R. L. Rodriguez, P. J. Greene, M. C.
Betlach, H. L. Heynekar, H. W. Boyer, J. H. Crosa, and
S. L. Falkow. 1977. Construction and characterization of
new cloning vehicles II. A multipurpose cloning system.
Gene. 2:95-113.

11. Chen, J. S., and L. E. Mortenson. 1974. Purification and
properties of hydrogenase from Clostridium pasteurianum
W5. Biochim. Biophys. Acta. 364:283-298.

12. Ching, Li., H. D. Peck, Jr., J. Legall, D. V.
Dervartanian, and A. E. Przybyla. 1987. Cloning and
sequence analysis of periplasmic hydrogenase from
Desulfovibrio gigas. Abstract. ASM Annual meeting,
Atlanta. p.216.

13. Chippaux, M., M. C. Pascal, and F. Casse. 1977. Formate
hydrogenlyase system in Salmonella typhimurium LT2. Eur.
J. Biochem. 72:149-155.

14. Close, T. J., D. Zaitlin, and C. I. Kado. 1984. Design
and development of amplifiable broad host-range cloning
vectors: analysis of the vir region of Agrobacterium
tumefaciens plasmid pTiC58. Plasmid. 12:111-118.

15. Cox, J. C., E. S. Edwards, and J. A. DeMoss. 1981.
Resolution of distinct selenium-containing format
dehydrogenase from Escherichia coli. J. Bacteriol.

16. Davis, R. W., D. Botstein, and J. R. Roth. 1980. A
manual for genetic engineering. Advanced bacterial
genetics. Cold Spring Harbor Laboratory, Cold Spring
Harbor, N. Y.

17. Ditta, G., S. Stanfield, D. Corbin, and D. R. Helinski.
1980. Broad host range DNA cloning system for
gram-negative bacteria: Construction of a gene-bank of
Rhizobium meliloti. Proc. Natl. Acad. Sci. U. S. A.
77: 7347-7351.

18. Evans, H. J., D. W. Emerich, T. Ruiz-Arqueso, R. J.
Meier, and S. L. Albrecht. 1980. Hydrogen metabolism in
the legume-rhizobium symbiosis, p.69-86. In W. H.
Orme-Johnson and W. E. Newton (ed.), Nitrogen fixation,
vol.11: symbiotic associations and cyanobacteria.
University Park Press, Baltimore.

19. Friedrich, B., C. Kortluke, C. Hogrefe, G. Eberz, B.
Siber, and J. Warrel-Mann. Genetics of hydrogenase from
aerobic lithoautotropic bacteria. Biochimie. 68:133-145.

20. Fukuyama, T., and E. J. Ordal. 1965. Induced
biosynthesis of formic hydrogenlyase in iron-deficient
cells of Escherichia coli. J. Bacteriol. 90:673-680.