(Received for publication, November 27, 1996, and in revised form, February 14, 1997)
From the Department of Pharmacological and Physiological Sciences, University of Chicago, Chicago, Illinois 60637
Mammalian adenylyl cyclases have two homologous
cytoplasmic domains (C1 and C2), and both
domains are required for the high enzymatic activity. Mutational and
genetic analyses of type I and soluble adenylyl cyclases suggest that
the C2 domain is catalytically active and the
C1 domain is not; the role of the C1 domain is to promote the catalytic activity of the C2 domain. Two
amino acid residues, Asn-1025 and Arg-1029 of type II adenylyl cyclase, are conserved among the C2 domains, but not among the
C1 domains, of adenylyl cyclases with 12 putative
transmembrane helices. Mutations at each amino acid residue alone
result in a 30-100-fold reduction in Kcat of
adenylyl cyclase. However, the same mutations do not affect the
Km for ATP, the half-maximal concentration (EC50) for the C2 domain of type II adenylyl
cyclase to associate with the C1 domain of type I adenylyl
cyclase and achieve maximal enzyme activity, or the EC50
for forskolin to maximally activate enzyme activity with or without
Gs. This indicates that the mutations at these two
residues do not cause gross structural alteration. Thus, these two
conserved amino acid residues appear to be crucial for catalysis, and
their absence from the C1 domains may account for its lack
of catalytic activity. Mutations at both amino acid residues together
result in a 3,000-fold reduction in Kcat of
adenylyl cyclase, suggesting that these two residues have additive
effects in catalysis. A second site suppressor of the Asn-1025 to Ser
mutant protein has been isolated. This suppressor has 17-fold higher
activity than the mutant and has a Pro-1015 to Ser mutation.
The activity of mammalian adenylyl cyclase, the enzyme that
converts ATP to cAMP, is the key step in modulating intracellular cAMP
concentration in response to extracellular stimulation by hormones,
neurotransmitters, and odorants. Nine isoforms of mammalian adenylyl
cyclases have been cloned to date, and they belong to a rapid expanding
cyclase family that includes class III adenylyl cyclases1 and guanylyl cyclases (1). All
nine isoforms have a similar structure, including two intensely
hydrophobic domains (M1 and M2) and two 40-kDa
cytoplasmic domains (C1 and C2). The two
cytoplasmic domains contain sequences (C1a and
C2a) that are homologous to each other and to class III
adenylyl cyclases and guanylyl cyclases. Each isoform of adenylyl
cyclase not only has distinct patterns of tissue expression but also
has unique responses to extracellular and intracellular stimuli (1-3).
In common, each adenylyl cyclase can be activated by the G
protein2 subunit, designated as
Gs
, and each can be inhibited by certain adenosine
analogues (P-site inhibitors). However, there are several
subtype-specific regulators, including forskolin, G protein
subunits, the
subunit of Gi, Go, and
Gz, Ca2+ ion, Ca2+-calmodulin,
Ca2+-calcineurin, cAMP-dependent protein
kinase, and PKC (for review, see Refs. 1-3). Certain ones of these
regulators (e.g.
) can be either stimulatory or
inhibitory, depending on the subtype of adenylyl cyclase (4). When
acting concurrently, the effects of regulators can be independent or
interdependent (antagonistic or synergistic). The binding sites
of adenylyl cyclases for Ca2+-calmodulin and G
protein
subunits are in the C1 and C2
domains, respectively (5-8). The sites for other regulators and the
mechanisms by which they modulate activity of adenylyl cyclase remain
to be determined.
One key step in studying the mechanisms for the complicated regulation
of adenylyl cyclase is to understand the molecular basis of catalysis.
The Gs- and forskolin-stimulated soluble adenylyl
cyclases can be formed either by mixing IC1 protein
(C1 domain of type I adenylyl cyclase) and IIC2
protein (C2 domain of type II adenylyl cyclase) in
vitro or by covalent linkage of IC1 and
IIC2 proteins (IC1IIC2) (9-12).
From the genetic and biochemical analysis of this soluble adenylyl
cyclase and mutational analysis of type I enzyme, we have shown that
both C1 and C2 domains are required for the
high enzymatic activity (13). Furthermore, the data suggests that the
C2 domain is the catalytic domain while the C1
domain enhances the activity of the C2 domain. In this paper, genetic and biochemical analyses are used to identify two amino
acid residues, asparagine and arginine that are crucial for catalysis
and might play an important role in inactivating the enzymatic activity
of the C1 domain.
Forskolin was from Calbiochem (La Jolla, CA);
2-d-3
-AMP was from Sigma; restriction enzymes and Vent DNA
polymerase were from New England Biolabs (Beverly, MA); Bradford
reagent was from Bio-Rad; ECL system was from Amersham Corp.; Ni-NTA
resin was from Qiagen (Chatsworth, CA); and Escherichia coli
XL1-red cells were from Stratagene (La Jolla, CA). GTP
S was
purchased from Pharmacia Biotech Inc. and purified by FPLC Mono Q
column chromatography.
E. coli strain, WJT-1
(cya AraD- endoA
crp/crp3
malT/malT3) was constructed
so that cAMP synthesis by the cloned cyclase replaced such synthesis by
the bacteria gene and that made it possible to enrich for mutations
that decreased synthesis of cAMP. To construct E. coli WJT-1
cells, E. coli TK2313 (F- thi rha lacZ
nagA
(kdpAE81) trk1 trkA405 endoA) was transduced with P1
lysate of a thr34::Tn10 strain, and a
tetracycline-resistant colony was purified. This strain was then
transduced to Thr+, araD139 with a
lysate of strain MC4100 (14). One araD139 derivative of
TK2313 was purified and transduced to
ilv500::Tn10 with a lysate of strain
MRi134 (pertinent marker is
ilv500::Tn10), and a
tetracycline-resistent colony was purified. The resulting strain was
transduced to ilv+,
cya with a
lysate of strain TK2648 (pertinent marker
(cya)8306, a
deletion originally obtained from J. Beckwith (Harvard Medical School)
in strain CA8306), and a
cya transductant was purified. Finally, the F141 plasmid that carried the crp malT TrkA
genes was introduced by selection for growth on medium containing 0.1 mM K+ (15).
Two conditions were used to select the loss of function
IC1IIC2 mutants in which expression of
cAMP-dependent genes was growth inhibitory or lethal. The
first was the sensitivity of E. coli to infection and
killing by phage, and the second was its inhibition of growth by a
metabolic intermediate of arabinose catabolism. A library of plasmids
that had mutations at the coding region of
IC1IIC2 was constructed by propagating plasmid
pProEx-HAH6-IC1IIC2 in E. coli
mutator strain, XL-1 red (mutD5 mutS mutT). Plasmid DNAs
were isolated and transformed into WJT-1 cells that harbored pBB131-Gs
Q227L [WJT-1 (Gs
Q227L)]. The
transformants were grown in Luria broth (LB) that contained 0.2%
maltose and supplement A (2 mM KCl, 5 mM
MgSO4, 100 µM IPTG, 25 µM
ampicillin, and 50 µM kanamycin) overnight. The
transformants (107 cells) were then infected with
phage,
vir (1010 plaque-forming units) in LB medium that
contained 1% maltose and supplement A for 1 h, and then plated on
LB plates that contained 1% maltose, 1% arabinose, and supplement A. Plasmid DNAs from the resulting colonies were isolated and transformed
into WJT-1 (Gs
Q227L). The resulting cells were tested on
MacConkey plates (MacConkey agar base, Difco) that contained 1%
maltose and supplement A. The coding sequences for
IC1IIC2 of the mutated DNAs from colonies that
were negative on maltose-MacConkey plates were then excised by
EcoRI and HindIII from plasmids encoding the
mutants and cloned into pProEx-HAH6 that had been digested with the
same enzymes. The DNAs that were transformed into strain WJT-1 cells
and were again negative on maltose-MacConkey plates were analyzed
further for expression and enzymatic activity of
IC1IIC2. Double-stranded DNA sequencing of the
coding region of IC1IIC2 mutants was performed to identify the mutations.
The second site suppressors from IC1IIC2-N1025S
mutants were obtained from E. coli that grew faster than the
control (E. coli that express both
IC1IIC2-N1025S and GsQ227L) on
maltose-minimal medium. To increase the sensitivity of cAMP detection,
E. coli WJT-3 cells (
cya cpdB
)
were constructed by transducing E. coli Crookes strain
(cpdB
) to
cya by the protocol
used to put this mutation into WJT-1 cells (16). The procedures in
analyzing the second site suppressors were similar to that in analyzing
the loss of function mutants of IC1IIC2
proteins.
To construct pProEx-HAH6-IIC2-M16 or
pProEx-HAH6-IIC2-N1025S(TCG), 0.8 kbp of DNA was amplified
by 15 cycles of polymerase chain reaction using
pProEx-HAH6-IC1IIC2-M16 or
pProEx-HAH6-IC1IIC2-N1025S(TCG) as the
template, respectively, Vent DNA polymerase, and two oligonucleotides (5-CGAGGAATTCTGGAGAACGTGCTTCCTGCACAC and
5
-TGCGTTCTGATTTAATCTGTATCAGGCTGA) as the
primers.4 The resulting DNAs were digested
with EcoRI and HindIII and ligated into
pProEx-HAH6 that had been digested with the same enzymes. The resulting
constructs were confirmed by dideoxy nucleotide sequencing. Kunkle's
method (17) was used to incorporate the desired point mutations into
the phagemid vectors, pProEx-HAH6-IIC2 and
pProEx-HAH6-IIC2-M16 (N1025S). Oligonucleotides (23-mer)
used for mutagenesis contained 10 complementary nucleotides flanking each side of the target codon, which was replaced with the appropriate codon for the mutants that had single point mutation (N1025S, N1025A,
N1025D, N1025K, N1025Q, N1025T, and R1029A) or double point mutations
(IIC2-P1015S, N1025S; P1015A,N1025S; P1015D,N1025S; P1015E,N1025S; P1015N,N1025S; and P1015Q,N1025S). Mutations were confirmed by dideoxy nucleotide sequencing of the plasmid DNA.
To construct pRC/CMV-ACII N1025K, the 0.7-kbp DNA was isolated from pProEx-HAH6-IIC2-N1025K that was digested with restriction enzymes, BspEI and XbaI, and cloned into pSK-ACII that had been digested with the same enzymes. The resulting plasmid was then digested with HindIII and XbaI, and the 3.3-kbp DNA, which encoded type II adenylyl cyclase, was excised and ligated to pRC/CMV that had been digested with HindIII and XbaI. The same approach was used to construct pRC/CMV-ACII for expression of type II adenylyl cyclase.
Expression and Purification of Wild Type and Mutant Forms of IIC2The plasmids that encoded wild type or mutant forms of IIC2 were transformed into E. coli BL21(DE3) cells. E. coli that harbored the desired plasmid were cultured in LB medium containing 50 mg/ml ampicillin at 30 °C. When A600 reached 0.4, IPTG (100 µM) was added. After 3-4 h, the induced cells were then collected and lysed, and IIC2 proteins were purified using the Ni-NTA column and FPLC Q-Sepharose column as described (9). The Coomassie Blue staining of SDS-polyacrylamide gel electrophoresis was used to determine the protein peak in the fractions from the Q-Sepharose column. The concentration of proteins was determined using Bradford reagent and bovine serum albumin as standard (18).
Expression of ACII-N1025K in HEK 293 (ATCC CRL1573) and Mouse Adrenal Y1Mouse adrenal Y1-t4 cells (kindly provided by Dr. E. Simpson, Univ. of Texas Southwestern Medical School) were selected for their high sensitivity to morphological changes upon cAMP stimulation. HEK 293 and mouse adrenal Y1-t4 cells (106/60-mm dish) were transfected with pRC/CMV, PRC/CMV-ACII, or pRC/CMV-ACII-N1025K by lipofection (Life Technologies, Inc). Geneticin (0.5 mg/ml) was used to select HEK 293 and mouse adrenal Y1-t4 cells that had stably incorporated the expression constructs, and cell colonies surviving antibiotic selection from a single transfection were pooled into a polyclonal population for subsequent analysis.
Adenylyl Cyclase AssayActivities of the soluble adenylyl
cyclases were assayed for 20 min at 30 °C in the presence of 10 mM MgCl2 (19). Wild-type IIC2 and
IIC2 mutant proteins were premixed with IC1 on
ice for 10 min before the assay. Recombinant Gs was
purified and activated by GTP
S as described (20, 21). Plasma
membranes from the pools of HEK 293 and mouse adrenal Y1-t4 stable
transformants were isolated as described (22). Enzyme activity of the
membrane fractions was assayed for 30 min at 30 °C in the presence
of 30 µM AlCl3, 10 mM
MgCl2, and 10 mM NaF
(AMF) or 100 µM forskolin (Fsk).
We have shown that
coexpression of IC1IIC2 and constitutively
active Gs, Gs
Q227L can complement the
genetic defects of E. coli
cya, which has a deletion of
the adenylyl cyclase gene and thus fails to synthesize cAMP (10). To
select IC1IIC2 mutants that have reduced
adenylyl cyclase activity (loss of function mutants), we have
constructed an E. coli strain, WJT-1 (
cya
AraD- endoA
crp/crp3
malT/malT3). Two conditions were used to
select the loss of function IC1IIC2 mutants in
which expression of cAMP-dependent genes was growth inhibitory or lethal. The first was the sensitivity of E. coli to infection by
phage. A virulent mutant of bacteriophage
(
vir) kills E. coli. Absorption of
to the cell
begins with binding to the LamB protein. Expression of LamB protein is
induced when E. coli is grown with medium containing
maltose, and expression of the LamB is dependent on cAMP. Thus, strains
that express the maltose genes are killed by
vir. Mutants that do
not express a functional LamB protein do not absorb
phage and thus
survive. Therefore, selection for resistance to
vir enriches for
mutants unable to make cAMP. As expected, most of WJT-1 cells
(>99.9%) that expressed both IC1IIC2 and
Gs
Q227L were sensitive to infection by
vir (plaques
were opaque), while those that expressed IC1IIC2 or Gs
Q227L alone were
insensitive to infection (data not shown).
The second condition for selection of loss of function mutants was the
growth inhibition of E. coli by a metabolic intermediate of
arabinose catabolism. Arabinose induces the enzymes specific to early
steps of arabinose catabolism. When ribulose-1-phosphate epimerase is
defective, the product produced by the araD gene, ribulose-1-phosphate, accumulates in the cell. This sugar phosphate, like many others, is growth inhibitory when accumulated to high levels.
Since expression of the genes for arabinose catabolism is also
dependent on cAMP, araD mutants that make cAMP are inhibited by arabinose, while those that do not make cAMP are not. E. coli WJT-1 strain was AraD, and thus 90%
of WJT-1 cells that expressed IC1IIC2 and
Gs
Q227L were growth-arrested when arabinose was added
(both in minimal and complex media). As a control, arabinose had no
effect on the growth of WJT-1 cells that expressed either
IC1IIC2 or Gs
Q227L alone. The
combination of selection by
vir and arabinose provided 104-fold enrichment of WTJ-1 cells that failed to produce
high concentrations of intracellular cAMP (Fig.
1A).
The first step in selecting IC1IIC2 mutants
that failed to be activated by GsQ227L in
vivo was random mutagenesis of the plasmid containing the coding
region of IC1IIC2 by growth in mutator strain
XL-1 red (mutT mutD mutS). This strain can introduce a
variety of mutations, including transitions, transversions, and
frameshifts. The cells that failed to produce cAMP were selected based
on their insensitivity to infection by
vir and to growth inhibition
by arabinose. Approximately 20-fold higher numbers of colonies of WJT-1
cells that carried the mutagenized plasmids (pProEx-HAH6-IC1IIC2) were observed than those
that carried the unmutagenized
pProEx-HAH6-IC1IIC2, indicating that this
selection did enrich for the loss of function mutants.
Thirty clones were selected for analysis. The coding regions for
IC1IIC2 of 30 colonies were introduced back to
unmutated pProEx-HAH6, and they were analyzed genetically (phenotype on maltose-MacConkey medium) and biochemically (immunoblot and adenylyl cyclase activity). Two IC1IIC2 mutants (M8 and
M16) were of interest. WJT-1 cells that carried either
IC1IIC2-M8 or M16 were insensitive to infection
by vir and growth arrest by arabinose and were delayed in being
positive on maltose-MacConkey medium (Fig. 1A). Lysates from
BL21(DE3) cells that expressed IC1IIC2-M8 and
M16 mutants had normal expression of 60 kDa
IC1IIC2, but they had a significant reduction
(4- and 95-fold, respectively) in Gs
- and
forskolin-stimulated adenylyl cyclase activity (Fig. 1, A
and B). IC1IIC2-M8 had a relatively
conserved leucine to phenylalanine change at aa 362 of type I adenylyl
cyclase (CTC
TTC) (Fig. 2).
IC1IIC2-M16 has two transitions, resulting in
the change of a highly conserved asparagine to serine at aa 1025 (AAC
AGC) and a silent mutation at aa 832 (TTG
TTA) of type II
adenylyl cyclase (Fig. 2).
Three IC1IIC2 mutants had 2-5-fold reduction
in expression of 60 kDa proteins that were correlated well with reduced
adenylyl cyclase activity (data not shown). Immunoblot showed that all three mutants had increased amounts of smaller molecular weight immunoreactive bands, suggesting that the reduced protein expression was associated with increased proteolysis (not shown). Sequence analysis of one of these mutants (M25) showed that
IC1IIC2-M25 had a glutamine to lysine change at
aa 848 of type II adenylyl cyclase (GAGAAG). We have shown that the
region from aa 821 to aa 855 of type II adenylyl cyclase is not
required for catalysis (9). Therefore, the mutation at aa 848 is likely
to cause a structural change that renders
IC1IIC2 highly susceptible to proteolysis. Although WJT-1 cells that carried the plasmids for the remaining 25 IC1IIC2 mutants did take longer to be positive
on maltose-MacConkey, E. coli lysates of these mutants did
not have a significant reduction in enzyme activity. Thus, they were
not analyzed further.
Genetic and mutational analyses suggest that the
C2 domain of mammalian adenylyl cyclase is the catalytic
domain, and the C1 domain is not (10, 13). This begs the
questions as to which amino acids are crucial for catalytic activity of
the C2 domains, and are they absent at the C1
domains? More than 30 different adenylyl cyclases and guanylyl cyclases
share the common homologous cytoplasmic domains, designated as the
cyclase domains (1). For the functional enzyme, the cyclase domains can
work as the homo-oligomer, such as adenylyl cyclases from yeast,
Trypanosome, Dictyostelium (ACG),
Leishmania, Brevibacterium, and
Rhizobium (23-30). The cyclase domains can also exist as an
obligatory heterodimer for enzymatic activity such as the
C1 and C2 domains from type I to type IX
adenylyl cyclases, Drosophila rutabaga, and
Dictyostelium ACA, and the and
subunits of soluble
guanylyl cyclases (22, 23, 31-46). The amino acid residues that are
crucial for catalysis should be conserved among all the cyclase domains
that function by themselves and among one of the two cyclase domains
among the obligatory hetero-dimers. Two amino acid residues (Asn-1025
and Arg-1029 of type II adenylyl cyclase) fits this criterion (Fig. 2,
letter b). They are conserved among all of the cyclase
domains that function by themselves. They are also conserved among the catalytic C2 domain of adenylyl cyclases and the
subunits of guanylyl cyclases but are not conserved in the
C1 domain of adenylyl cyclases or the
subunit of
soluble guanylyl cyclases. We thus examined the effects of mutations at
Asn-1025 and Arg-1029 on enzyme activity of
IC1+IIC2 complex.
To evaluate the effects of mutations of Asn-1025 of type II adenylyl
cyclase, we constructed and purified IIC2 mutant proteins that had the Asn-1025 changed to Thr, Ser, Ala, Gln, Asp, and Lys
(IIC2-N1025T, N1025S, N1025A, N1025Q, N1025D, and N1025K). All six mutants expressed normal amounts of enzyme compared with wild-type IIC2 based on immunoblot of E. coli
BL21(DE3) lysates (Fig. 3). They also could be purified
readily (Fig. 3). The biochemical properties of these
IIC2-N1025 mutants are summarized in Table I
and Fig. 4. All the enzyme assays were performed in the
presence of a fixed amount of IC1. Mutations of Asn-1025
caused a range of reduction (10-5000-fold) in forskolin- and
Gs-activated adenylyl cyclase activity
(Kcat) with the order of N1025T, N1025S, N1025A,
N1025Q, N1025D, and N1025K from the highest enzyme activity to the
lowest. IIC2-N1025T and IIC2-N1025S had similar
values of Km for ATP and of half-maximal
concentration (EC50) for forskolin stimulation in the
presence of GTP
S-Gs
compared with wild-type
IIC2. In addition, similar values were observed in
EC50 for wild-type IIC2 and IIC2
mutants (N1025T and N1025S) to form complexes with IC1 and
reach maximal enzyme activity that was stimulated by either forskolin
alone or forskolin plus GTP
S-Gs
. Although the enzyme
activity of mutants IIC2-N1025A, N1025Q, N1025D, and N1025K
was too low to determine Km values of ATP and
EC50 values for forskolin stimulation and for
IIC2 requirement, all four mutant proteins could
effectively compete with wild-type IIC2 to form a complex
with IC1 and block wild-type IIC2 enzyme activity (Fig. 4D). The concentrations of IIC2
mutants required to achieve half-maximal inhibition (IC50)
were close to the concentration of wild-type IIC2 protein
in the assays (0.26 µM). This result indicated that the
mutation of Asn-1025 caused reduction in catalysis without gross
structural alteration. Both IIC2 mutants (N1025T and
N1025S) had reduced sensitivity to 2
-d-3
-AMP, an inhibitor of adenylyl cyclase activity, when both forskolin and
GTP
S-Gs
were present (Fig. 3C and Table
I).
|
We then examined the effects of the mutation from arginine to alanine
at aa 1029 (IIC2-R1029A). IIC2-R1029A could be
expressed normally and was readily purified to homogeneity (Fig. 3).
The biochemical properties of IIC2-R1029A are similar to
IIC2-N1025S (Table I and Fig. 4). When mixed with
IC1, IIC2-R1029A had 30-fold reduction in
catalysis compared with wild-type IIC2. Similar values to
that of wild-type IIC2 were observed in
Km of ATP, EC50 for forskolin to
maximally activate the complex of IC1 and IIC2-R1029A, and EC50 for
IIC2-R1029A to form a complex with IC1 for
reaching maximal activation (Table I and Fig. 4). The IC50 value of 2-d-3
-AMP for IIC2-R1029A was about
200-fold higher than that of wild-type IIC2. A similar
mutation in recombinant type I adenylyl cyclase was analyzed and this
mutation also had reduced catalysis without altered ability of the
mutant protein to interact with GTP
S-Gs
(13).
We also examined the effects of combined mutations at both arginine 1029 to alanine and asparagine 1025 to serine (IIC2-N1025S,R1029A). IIC2-N1025S,R1029A could be expressed and purified similar to wild-type IIC2 (Fig. 3). It had 3,000-fold reduction in activity compared with wild-type IIC2 (Table I). When mixed with wild-type IC1 and IIC2 proteins, IIC2-N1025S,R1029A was a potent inhibitor for enzyme activity, indicating that it interacted with IC1 normally. Together, this result indicates that both Asn-1025 and Arg-1029 are crucial for catalysis, and the effect of mutations on catalysis by these two residues is additive.
Effect of Asparagine Mutation at aa 1025 on Type II Adenylyl CyclaseTo evaluate the effect of Asn-1025 on full-length type II
adenylyl cyclase, we constructed expression vectors with the
cytomegalovirus (CMV) promoter that carried cDNA for type II
adenylyl cyclase with or without the N1025K mutation. HEK 293 and mouse
adrenal Y1-t4 cells were transfected with the expression vectors, and 50-200 colonies of stable transformants were pooled together and analyzed for their adenylyl cyclase activity (Fig. 5).
As expected, plasma membranes from HEK 293 and mouse adrenal Y1-t4
cells that had stably integrated the vector with type II adenylyl
cyclase had 2-8-fold increases in adenylyl cyclase activity over cells that had integrated the control vector. When stimulated by forskolin, plasma membranes from HEK 293 and mouse adrenal Y1-t4 cells that had
integrated the vector with type II adenylyl cyclase containing the
N1025K mutation not only did not have increased enzyme activity but had
a 25-50% reduction in enzyme activity compared with that from the
control cells. This indicates that type II adenylyl cyclase containing
the N1025K mutation is catalytically inactive and can dominantly block
the activation of endogenous adenylyl cyclases. The mechanism for such
blocking is currently under investigation.
Selection and Analysis of the Second Site Suppressor of the IC1IIC2-N1025S
Cyclic nucleotide
phosphodiesterase hydrolyzes cAMP, and thus a defect in the
cpdB gene enhances accumulation of intracellular cAMP.
Therefore, an E. coli strain, WJT-3 (cya
CpdB-), was constructed to increase the
sensitivity of our genetic screen. As expected, E. coli
WJT-3 cells can grow on maltose-minimal medium with 10-fold less cAMP
than is required by E. coli
cya, CpdB+. When harboring plasmids for expression of
Gs
Q227L and IC1IIC2, E. coli WJT-3 cells can grow about twice as fast (2 days) on
maltose-minimal medium than E. coli
cya,
CpdB+ (4 days).
We employed the second site suppressor approach to search for amino
acid residues that can suppress the catalytic defect of IC1IIC2-N1025S. A library of randomly mutated
plasmids that encoded IC1IIC2-M16 was generated
by passing through E. coli XL-1 red cells and then
transformed to E. coli WJT-3 cells that expressed Gs-Q227L. More than 200 colonies that carried mutated
IC1IIC2-M16 and Gs
-Q227L were
grown within 3 days, similar to those that carried wild-type
IC1IIC2 and Gs
-Q227L. Eight
independent clones that had different colony sizes were selected to
further analyze. All eight IC1IIC2 revertants
had similar adenylyl cyclase activity and time required for being
positive on maltose-MacConkey, as did wild-type
IC1IIC2. Sequence analysis revealed that all of
them had a transition to revert serine back to asparagine at aa 1025. This result indicates again that asparagine at aa 1025 is crucial for
the enzyme activity of adenylyl cyclase.
To avoid the true revertants in the genetic screen, we constructed
pProExHAH6-IC1IIC2-N1025S(TCG) which had the
TCG as codon for serine at aa 1025. In this plasmid, three mutational
events are required to revert the TCG codon to the codon for asparagine (AAC/T). A library of randomly mutated
pProEx-HAH6-IC1IIC2(TCG) plasmids was generated
by passage through E. coli XL1 red cells and was used to
select the revertants as described above. As expected, E. coli WJT-3 cells that expressed IC1IIC2-N1025S and
GsQ227L were negative on maltose-minimal medium and were
not positive on maltose-MacConkey plates (Fig. 6). When
coexpressed with Gs
-Q227L in E. coli WJT-3
cells, two independent IC1IIC2 revertants
allowed E. coli to grow in 4 days on maltose-minimal medium
and were positive within 44 h on maltose-MacConkey plates,
significantly better than the IC1IIC2 mutant
but not as well as wild-type IC1IIC2 (Fig. 6A). Lysates from E. coli BL21(DE3) cells that
expressed IC1IIC2 revertants had normal
expression of 60 kDa IC1IIC2. Consistent with
the results from the genetic analysis, lysates from E. coli BL21(DE3) cells that expressed IC1IIC2
revertants had 17-fold higher adenylyl cyclase activity than
IC1IIC2 mutants but 10-fold lower than
wild-type IC1IIC2 (Fig. 6A). DNA
sequencing analysis revealed that both revertants retained the TCG
codon at aa 1025 but had acquired a proline to serine at aa 1015 (CCA
TCA) of type II adenylyl cyclase.
Characterization of IIC2-P1015,N1025S) Mutations
To evaluate the effect of proline mutation at aa 1015 of type II enzyme, we constructed six IIC2 mutants, all
having serine at aa 1025 and a mutation from Pro 1015 to Ser, Ala, Asp,
Glu, Asn, or Gln (IIC2-P1015S,N1025S; P1015A,N1025S;
P1015D,1025S; P1015E,N1025S; P1015N,N1025S; and P1015Q,N1025S).
Based on immunoblot analysis of lysates from E. coli
BL21(DE3) cells, IIC2-P1015S,N1025S, IIC2-P1015A,N1025S, and IIC2-P1015E,N1025S were
expressed normally, whereas IIC2-P1015D,N1025S,
IIC2-P1015N, N1025S and IIC2-P1015Q,N1025S had 50-90% reductions in expression as the soluble proteins. To characterize these proteins, all were purified to near homogeneity (Fig. 6B). The biochemical properties of these mutants
are summarized in Table II. All six mutants had
3-10-fold higher Gs- or forskolinstimulated enzyme
activities compared with IIC2-N1025S, while their enzyme
activities were still significantly lower than wild-type
IIC2. The IC50 values for
2
-d-3
-AMP to inhibit the enzyme activity were slightly
reduced compared with IIC2-N1025S (particularly
IIC2-P1015D,N1025S and IIC2-P1015N,N1025S) but
were still significantly higher than that of wild-type
IIC2. Their Km values for ATP and
EC50 values for IIC2 mutants to achieve the
maximal forskolin-stimulated enzyme activity were similar to that of
wild-type IIC2. Interestingly, the EC50 values for forskolin to maximally activate the complex of IC1 and
IIC2 mutants were altered in some mutants
IIC2-P1015S,N1025S; IIC2-P1015A,N1025S; IIC2-P1015N, N1025S; and
IIC2-P1015Q,N1025S).
|
Mammalian adenylyl cyclase has a complex molecular design,
presumably to integrate and respond to various extracellular and intracellular stimuli. Genetic and biochemical analyses of type I and
soluble adenylyl cyclases indicate that both the C1 and C2 domains are essential for the high enzyme activity and
only the C2 domain is
catalytic.5 In this paper, we have shown
that the conserved asparagine and arginine residues are crucial for
catalysis, and they might be two of the key residues that prevent the
C1 domain from being catalytically
active.6 The residue in the C1
domain that corresponds to Asn-1025 of the C2 domain is
commonly a threonine. Interestingly, this replacement of Asn-1025 in
the C2 domain results in the retention of the highest activity of the replacements examined (Fig. 2). The corresponding residue of the conserved C2 arginine in the C1
domain is either histidine or lysine (Fig. 2). We have shown that the
arginine to lysine substitution in type I adenylyl cyclase did not
alter the enzyme activity, highlighting the role of the substitution at
the corresponding site of the C2 asparagine in inactivating the enzyme activity of the C1 domain (13). The role of the
conserved asparagine and arginine in catalysis remains to be
determined. One possible role is to coordinate and stabilize the
PO5 transition state similar to the way in which the Gln
and Arg are involved in the hydrolysis of GTP in the G protein subunit (47, 48). Structural determination of the cyclase domain with
its substrate should help to elucidate the roles of these two residues
in catalysis.
The changes in free energy (G) have been used to
determine whether the two mutations have an additive effect on
catalysis (49). If the free energy change due to the first mutation is denoted by
Gx and that due to the second
mutation is
Gy, the change in free energy
observed in the double mutant,
Gx,y, can be expressed as
Gx,y =
Gx +
Gy +
Gi, where
Gi
represents an interaction energy between sites. When
Gi is near 0, the two sites behave
independently, and the double mutant displays "simple additivity."
The change in transition-state stabilization energy (
GT) is equal to
RT*ln(Kcat/Km)mutant/(Kcat/Km)wild-type. Thus, both
GT for both mutant
IIC2-N1025S and IIC2-R1029A are 2.2 kcal/mol.
We could not determine Km of double mutant IIC2-N1025S,R1029A due to its extremely low enzyme
activity. However, it might be reasonable to assume that the
Km of the double mutants is not altered
significantly because we observed no change in Km
for either IIC2-N1025S or IIC2-R1029A mutants. If this is true,
GT for double mutant
IIC2-N1025S,R1029A would be 4.8 kcal/mol, close to the sum
of
GT (4.4 kcal/mol) for mutant IIC2-N1025S and IIC2-R1029A. Thus, it appears
that the conserved Asn-1025 and Arg-1029 act independently in the
catalytic activity of adenylyl cyclase.
How does a proline to serine mutation at aa 1015 increase 10-15-fold
enzyme activity of IIC2-N1025S? Proline can exist as either
cis- or trans-peptide bonds. However, a charged lysine residue, not
Phe, Tyr, or Leu, immediately proceeds proline at aa 1015, making it
less likely to be cis-Pro (50) (Fig. 2). The entropy effect of proline
contributes positively in protein stability; thus, mutation at a
proline residue might introduce an increased flexibility (50, 51). We
hypothesize that the open-close conformational switch is essential for
the catalytic activity of adenylyl cyclase and that the increased
flexibility of the protein by the proline to serine mutation might
facilitate such a switch (1, 13). Further experiments are required to test this hypothesis. Proline is involved in numerous gain of function
mutants in other proteins. Proline to leucine has been shown to render
G protein-coupled -factor receptor to be constitutively active and
second site suppressor mutations from other amino acids to proline
enhance the functioning of several mutated proteins, i.e.
triosephosphate isomerase, F1F0 ATP synthase,
and lactose permease (52-55).
Three genetic model systems have been set up to study catalysis and regulation of adenylyl cyclases that have 12-transmembrane helices (10, 56-58). In both E. coli and Saccharomyces cerevisiae, the paradigm of cAMP-dependent cell growth under certain growth conditions are utilized, while in D. discoideum, that of cAMP-mediated cell aggregation is applied. Conditions that allow selection of mutants that gain or lose cAMP accumulation have been established both in E. coli and in D. discoideum. Mutations that inactivate or increase the basal enzyme activity have been found in ACA and in soluble adenylyl cyclase (56, 57, see above). These systems should permit the genetic analysis of interactions between adenylyl cyclases and its regulators. In addition, these systems should help us to identify the pharmacological reagents that could modulate the enzyme activity of adenylyl cyclase with isoform specificity.
We thank David Hahn for superb technical help; W. Epstein for helpful suggestions in constructing E. coli WJT-1 and WJT-3 cells and the genetic screening; and W. Epstein, H. Fozzard, and C. Skoczylas for critical reading of the manuscript.