Download Localization of Protein-Protein lnteractions between Subunits of

The Plant Cell, Vol. 4, 161-171, February 1992 O 1992 American Society of Plant Physiologists
Localization of Protein-Protein lnteractions between
Subunits of Phytochrome
Michael D. Edgerton and Alan M. Jones’
Department of Biology, University of North Carolina, Chapel Hill, North Carolina 27599-3280
We have used a nove1assay based on protein fusions with h repressor to identify two small regions within phytochrome’s
carboxy-terminal domain that are capable of mediating dimerization. Using an in vivo assay, fusions between the DNA
binding, amino-terminal domain of h repressor and fragments from oat PhyA phytochrome have been assayed for increased repressor activity, an indicator of dimerization. In this assay system, regions of oat phytochrome between amino
acids V6234673 and N1049-Q1129 have been shown to increase repressor activity. These short spans are highly conserved between proteins belonging to the phytochrome PhyA family. Embedded within these sequences are four segments
that could potentially form amphipathic a helices. Two of the segments are well conserved between PhyA phytochrome
and phytochromes encoded by the phyB and phyC genes, suggesting that heterodimers might form by way of subunit
interaction at these sites.
INTRODUCTION
Phytochrome is a plant photoreceptor involved in photomorphogenic events ranging from seed germination and
deetiolation to the induction of flowering. Phytochrome initiates some of these diverse responses over a fluence range
that spans at least seven orders of magnitude (e.g., Jones et
al., 1991). In this regard, phytochrome is unique among known
photoreceptors. Also, phytochrome is uniquely photoreversible between a red light-absorbing form (Pr) and a far-redlight-absorbing form (Pfr) using red or far-red light, respectively. The Pfr form is thought to be the physiologically active
form of phytochrome. In the natural environment, photoreversibility enables plants to detect both lightldark transitions and
changes in redlfar red light ratios that occur under conditions
such as shading in a forest canopy (Smith, 1982).
Phytochrome is a homodimer of two independently photoreversible subunits. Each subunit is a 120- to 127-kD
polypeptide with a covalently linked chromophore. Limited proteolysis has demonstratedthat phytochrome is divided into two
major domains. The protein is composed of (1) a 65-kD aminoterminal domain that includes the chromophore and all structural requisites necessary for photoreversibility (Jones et al.,
1985) and (2) a 60-kD carboxy-terminal domain that contains
structures required for dimerization (Jones and Quail, 1986)
and sequences possibly involved in rapid, light-mediated
To whom correspondence should be addressed at Department of
Biology, 010A Coker Hall, CB# 3280, University of North Carolina,
Chapel Hill, NC 27599-3280.
degradation of the protein (Shanklin et al., 1989). Phytochrome
quaternary structure has been studied using ultracentrifugation (Jones and Quail, 1986), electron microscopy (Jones and
Erickson, 1989), and size exclusion chromatography (SEC)
(Lagarias and Mercurio, 1985).These studies and others have
produced a rough picture of phytochrome quaternary structure. Phytochrome is tripartite in structure, consisting of two
interacting carboxy domains that dimerize to form a base from
which extend the two globular amino-terminal domains. The
amino- and carboxy-terminal domains of each subunit are
joined by a region defined as the hinge (Jones and Erickson,
1989).
Comparison of the known physical properties of phytochrome with kinetic descriptions of phytochrome responses
has led to a dimeric model for phytochrome action (VanDerWoude, 1985,1987). In VanDerWoude’s model, cells are thought
to preceive mixed dimers in which one subunit is in the Pr form
and the second subunit in Pfr conformation at low light levels.
A second phytochrome-responsivesystem would also be measuring total Pfr levels at higher light levels. This model provides
a possible explanation for the extraordinary fluence sensitivity of this photoreceptor and emphasizes the importance of
its dimeric structure in attaining this sensitivity. Also, the existente of two or more different types of phytochrome raises the
possibility of heterodimer formation. Pratt et al. (1991) have recently provided biochemical evidence that such heterodimers
do indeed form in solution. For these reasons, we have focused
our efforts on identification and characterization of regions of
the protein involved in dimerization.
162
The Plant Cell
We have developed a nove1 assay based on gene fusions
to the DNA binding domain of the h repressor (cl) that has
enabled us to identify segments of phytochrome capable of
mediating homotypic protein-protein interactions that may be
involved in dimerization.
RESULTS
h Repressor Assay for Dimerization
Covalent cross-linking and SEC have been used successfully
to map the dimerization region of phytochrometo a 37-kDtryptic
fragment derived from the carboxy-terminalhalf of phytochrome
(Jones and Quail, 1986). However, these methods have been
limited in their ability to more precisely localize the dimerization region because of (1) aggregation of smaller phytochrome
fragments (Yamamoto and Tokutomi, 1989), (2) the low resolution to which most monoclonal antibody epitopes have been
mapped (Cordonnier, 1989), and (3) the limitations of SEC in
distinguishing between dimeric molecules and molecules with
extended shapes.
We have developed an assay for dimerization that allows
fine mapping of dimerization regions or, more precisely, sites
of homotypic protein-protein interaction, and avoids the problems mentioned above. The premise of our assay is to begin
with a protein that must be a dimer to function, remove that
protein's dimerization region, and replace it with segments of
a second protein. If one of the segments of the second protein
contains a region through which the new chimeric protein is
able to properly self-associate, activity will be restored.
We have used the h repressor (cl) as the first component
in our assay system. The native h repressor is a homodimer
of 26-kD (236 residues) subunits. The protein is divided into
two domains: an amino-terminal, DNA binding domain
(residues 1 to 92) and a carboxy-terminal domain (residues
132to 236), which includes almost all of the structures necessary for dimerization of the repressor. The native h repressor
binds to its operator sequence (OR1),with a dissociation constant of -2 x 10-8 M. In contrast, h repressor's aminoterminal domain, which contains only a single interprotein
contact site, binds to OR1with a dissociation constant of -6
x
M (Sauer et al., 1990). In our h repressor fusion vector, pME10, we have placed a polylinker immediately following
the sequence encoding repressor helix 5, which ends at V92.
Helix 5 was included in this construct even though it is a weak
dimerization region because it may be requiredfor proper positioning of two amino-terminal domains on their operator
sequence (Hecht et al., 1983; Weiss et al., 1987).
In our dimerization assay, illustrated schematically in Figure 1, h repressor fusions are expressed under the control of
the lactose promoter in pME10. Levels of expression from this
promoter are increased by the addition of isopropyl P-D-thiogalactopyranoside (IPTG), which reduces the affinity of Lacl
for its operator sequence. To report repressor activity,
f
LacZ
hPf7
Figure 1. Schematic Diagram of h Repressor Dimerization Assay.
Expression of 1 repressor fusion proteins from pMElO constructs is
controlled by a lac promoter. The addition of IPTG leads to increased
levels of fusion protein expression through its interaction with Lacl.
To report repressor activity, expression was carried out in a cell line
that contains lacZ under the control of ~ P RUpon
.
dimerization, h
repressor fusions bound to PR,decreasing p-galactosidasesynthesis. Thus, the amount of p-galactosidaseactivity observed ata given
leve1of IPTG depended upon the ability of the h repressor fusion proteins to form active dimers. Amp', ampicillin resistant; aa, amino acid.
expression is carried out in a strain that contains a lacZ gene
under the control of the h rightward promoter (PR). Thus, in
this system, addition of IFTG leads to a decrease in fi-galactosidase levels. The manner in which these levels decrease
is dependent upon both the affinity with which the repressor
binds to its operator and the amount of repressor present.
As an initial test of this system, we replaced the carboxyterminal half of h repressor with a portion of a yeast regulatory protein, GCN4, known to contain a leucine zipper
dimerization motif (Hope and Struhl, 1987). The plasmid expressing this protein, pMElO/GCN4, contains sequences
encoding GCN4 amino acids L169-R281 inserted into the Smal
site of pME10. The resulting protein was stable in Escherichia
coli, accumulating to 1 to 2% of total protein when induced
with 0.5 mM IPTG. Cells containing pMElOlGCN4 were resistant to h phage infection. PGalactosidase levels measured in
cells lysogenic for h200 were similar to those expressing
the MElO/GCN4 protein and those expressing full-length h
Phytochrome Dimerization
represser, as shown in Figure 2A. Furthermore, the protein
could be seen to form dinners when assayed by SEC, as shown
in Figure 2B. As estimated by size exclusion data, the ME10/
GCN4 fusion protein has a dimerization constant in the micromolar range (between 2.5 and 25 nM). Hu et al. (1990) have
independently used a similar assay to further study the leucine zipper region of GCN4. Their assay, as did ours, had
approximately a 10-fold range in sensitivity.
PLASMID
#pfu
[IPTG]x10' 6 M
[IPTGJx 10-6M
0
1.0
10
0
1.0
10
950
850
750
0
0
0
4700
4900
2500
132
122
102
450
250
200
0
0
0
FG750
ME10
(5-Galactosidase
ME10/GCN4
B
Fraction Number
5
10
15
20
-•-71
— 44
3
8)
en
-28
ME10/_
GCN4
(0
-18.3
-15.3
t
t
67 45
t
29
Apparent Molecular Mass (kD)
Figure 2. A ME10 Fusion Protein Containing a Leucine Zipper Behaves as a Dimer.
(A) A fusion between the amino-terminal DNA binding domain of X
represser and sequences including a leucine zipper dimerization motif from GCN4 (L169-R281) was able to prevent lytic growth of wild-type
X phage and suppress p-galactosidase activity in a manner similar
to full-length X represser (encoded by FG750) when induced with IPTG.
(B) ME10/GCN4 fusion protein was analyzed by SEC and found to be
in equilibrium between monomeric and dimeric forms. Bacterial extract (100 jiL) containing ~25 mM ME10/GCN4 was separated on a
TSK2000 size exclusion column. Individual fractions were collected
and analyzed by immunoblotting with anti-X represser antiserum NCOS.
Elution points for carbonic anhydrase (29 kD), ovalbumin (45 kD), and
BSA (67 kD) are indicated. Fusion protein can be seen to have eluted
as two peaks, presumably representing monomeric and dimeric forms.
Both peaks migrated with slightly larger molecular mass than would
be expected if ME10/GCN4 (213 residues) were a simple globular protein (48 kD expected for the dimer, 24 kD for the monomer),
pfu, plaque-forming units.
163
Potential Dimerization Regions Reside between
Residues V623-S673 and N1049-Q1129
Our efforts to map potential dimerization region(s) in phytochrome have focused on the carboxy-terminal half of the protein
that has previously been shown to behave as a dimer (Jones
and Quail, 1986). We have defined the carboxy-terminal domain as beginning at approximately amino acid L600 of oat
PhyA APS, based on sequence data derived from fragments
of proteolytically digested phytochrome (Grimm et al., 1988).
Throughout this report, phytochrome amino acid designations
will be in reference to those encoded by the oat gene phyA
X\P3 (Hershey et al., 1985), with the initiation methioninecodon
encoding amino acid 1.
In reporting the results from our in vivo assay, we have defined a new constant, relative repression (Rr), which is a
unitless value representing the relative ability of each fusion
to repress p-galactosidase production. This value is derived
by dividing the concentration of IPTG yielding half-maximal
repression for a given construct by the concentration of IPTG
that yields half-maximal repression for pME10 with no insert.
Because the dimerization constant for phytochrome has not
been measurable using conventional dissociation techniques
(e.g., Jones and Quail, 1986; Choi et al., 1990), we use the
Rr value as a relative measure of each fusion protein's dimerization coefficient.
Segments encoding pieces of the carboxy-terminal half of
oat phytochrome were excised from pCIB315 (Thompson et
al., 1989) and inserted into pME10. Initially, an Xmnl fragment
encoding amino acids V623-E1048, nearly all of the carboxyterminal half, was ligated into the Smal site of pME10. As
expected, this construct, pME10/PC623-1048, repressed
(3-galactosidase activity as IPTG levels increased. pME10/
PC623-1048 showed half-maximal repression of p-galactosidase at 0.9 ± 0.4 uM IPTG (n = 4). pME10 showed halfmaximal repression at 5.4 ± 0.1 uM IPTG (n = 4), yielding
an Rr value of 6 ± 0.5 for pME10/PC623-1048. Thus, the level
of IPTG giving half-maximal repression was roughly sixfold
lower than the truncated X represser, ME10, encoded by pME10,
as shown in Figure 3A. Unlike ME10/GCN4, the protein encoded by this construct and most other ME10/phytochrome
fusions did not accumulate to significant levels when induced
with high levels of IPTG (0.5 mM).
Smaller fragments of phytochrome coding sequences spanning the entire carboxy-terminal domain were assayed
individually (Figure 3B). Only four of the fusions were found
to have Rr values significantly higher than pME10. Three constructs, pME10/PC488-673, pME10/PC623-723, and pME10/
PC599-683, all have in common sequences encoding a 5.5-kD
region located between residues V623 and S673. Other constructs lacking these sequences had represser activity
approximately equal to that encoded by pME10 (Figure 3A).
One other construction with a relatively high Rr value was
pME10/PC914-1129. Because represser fusion proteins
containing the phytochrome segment between M914 and
E1048 do not have represser activity, we concluded that a
164
The Plant Cell
To separate these two factors, we have measured levels of fusion protein produced upon induction with 100 \iM IPTG (the
lowest level of IPTG needed to reliably detect the fusion protein in immunoblots of crude extracts) by five constructs and
pME10 by immunoblotting with polyclonal antiserum, NCOS,
raised against X. represser. As shown in Figure 4, protein concentration was found to vary between constructs when induced
to this level; however, this variation did not correlate with Rr
values. For example, plasmids encoding proteins with Rr
values of >1.0 were seen to be expressed at both relatively
high (pME10/PC488-673) and low (pME10/PC599-683) levels.
Plasmids encoding fusion proteins with Rr values of approximately 1.0 were also seen to vary in abundance. Thus, the
observed Rr values do not reflect simple changes in either
protein abundance or represser strength alone.
CO.
1.0
2.5
5.0
10
25
50
[IPTG] (uM)
B
Amino Domain
• CO2H
ME10/PC623-1048
623———————
ME10/PC623-723
623 iri-i——723
ME10/PC488-673 488-
ME10/PC599-683
3
4
5
6
Rr
-1048
6.0
-11048
0.7
45.2
4.3
^673
ME10/PC849-982
ME10/PC488-557
2
2.6
H:;:>723-
ME10/PC723-1048
1
Carboxy Domain
849-
-982;i;
488——557;g
0.5
1
599-SS-683
ME10/PC685-815
(585———815
ME10/PC914-1129
ill
0.8
2.2
0.9
914-
(0
CO
29.5
2.4
Figure 3. Phytochrome Dimerization Regions Are Located between
Residues V623-S673 and N1049-Q1129.
(A) Dose-response curves for ME10/phytochrome gene fusions. Gene
fusions between X represser and phytochrome segments were expressed in cells lysogenic for X200 and assayed for 3-galactosidase
activity as IPTG concentration was increased. ME10 (•), ME10/PC6231048 (A), ME10/PC623-723(O), ME10/PC685-815 (D), and FG750 (fulllength c/) (•). The abscissa is log scale.
(B) Map of ME10/phytochrome fusion proteins and their relative repression values. Relative repression values (Rr) are defined as the
concentration of IPTG necessary to half maximally repress 3-galactosidase synthesis by ME10 divided by the concentration of IPTG, which
half-maximally represses p-galactosidase expression with the construct
being assayed. R, values are the mean of at least three separate experiments. All amino acid numbers are in reference to the oat phyA
AP3 gene (Hershey et al., 1985).
dimerization structure in the protein ME10/PC914-1129 resides
carboxyl to residue E1048. Therefore, these data indicated that
potential dimerization regions are located between residues
V623-S673 and N1049-Q1129.
Relative repression values are a function of both the dimerization constant and protein concentration for each construct.
JS
3
U
«
O
18.7
15.5
Rr
1.0
0.9 2.6 0.7 4.3 0.8
Figure 4. Expression Levels of ME10 Fusion Proteins.
Soluble protein (100 |ig) from cells induced with 100 uM IPTG was assayed for fusion protein expression levels by immunoblotting with anti-X.
represser antiserum NCOS. The relative repression value (Rr) for each
fusion is displayed below each lane. Lane 1, ME10; lane 2, ME10/PC685815; lane 3,10/PC599-683; lane 4, 10/PC723-1048; lane 5, 10/PC488673; lane 6, ME10/PC849-982. Levels of individual fusion protein can
be seen to vary when induced with 100 uM IPTG. There is no apparent correlation between protein concentration and R, values at
this level of induction. In each lane, two nonspecific bands can be
seen; these serve as serendipitous internal controls for total protein
concentration.
Phytochrome Dimerization
To further address the relative contributions of protein concentration and DNA binding affinity to R, values, we have
used a filter binding assay to measure the amount of active
DNA binding protein present in extracts of cells grown in the
concentrations of IPTG used in measurements of R, values.
PGalactosidase level was also determined for each IPTG concentration by both immunoblotting of the crude extracts and
by activity measurement. By using saturating levels of radiolabeled operator DNA, we are able to measure the total
abundance of OR1binding protein, independent of binding affinity (Riggs et al., 1970). Figure 5 shows that at the low levels
of IPTG required to repress lacZ, the amount of DNA binding
activity in extracts expressing MElO/PC914-1129 fusion protein and ME10 protein is the same, indicating that the level
of the phytochrome repressor fusion protein and MElO protein is the same at the IPTG concentrationsmeasured. However,
the repressivity of ME101PC914-1129 protein, as measured
through p-galactosidaselevels, is much greater than for ME10
protein. Therefore, the decrease in p-galactosidase activity
reported in R, values for pMElO/PC914-1129 is due to an
increase in the affinity of the fusion protein for operator sequences dueto an improved ability to dimerize and not due
to a difference in protein levels.
DlSCUSSlON
We are interested in the protein structure that forms and maintains phytochrome as a dimer because this dimeric character
112.
3OW
I
t
'
1
"""I
' ' '
""'I
-i 1.50
-1.25
7
I
I
-"
- 0.75 2
a
-0.50
5w
t
0 9 Av
..,::
.-,.--_.
. . ,.,
#.
e-Ya
12:5'
- 0.25
6
O"
165
influences phytochrome fluence sensitivity and possibly its
range of different responses.
We began our study by noting that phytochrome behaves
as a dimer even at concentrations at the lower detection limits
of analytical assays and that phytochrome secondary structure is more sensitive to denaturants than its quaternary
structure, suggesting to us that standard equilibria methodologies would not work to characterize dimerization (A.M. Jones,
unpublished data). lnitial experiments with phytochrome expression in nonfusion systems were not successful because
nearly all phytochrome sequences we expressed were nonnative, although coexpression with chaperonins did help
significantly (M.D. Edgerton and A.M. Jones, unpublished
data). It was therefore necessary to develop a new method
for probing phytochrome structure. The method we have developed relies on modifying a stable prokaryotic protein, hcl,
such that its activity is dependent, in part, upon additional foreign protein sequences involved in dimerization.
Using this technique, we have identified two regions of
phytochrome that are involved in dimerization, one residing
between residues V623-S673 and the other between N1049Q1129. Each region when fused to the truncated h repressor
restored repressor activity. In the first region, V623-S673, severa1 overlapping clones were assayed together with flanking
regions to delineate the dimerization region. We have used
immunoblot analysis to show that while the concentrations of
these proteins vary at high IPTG levels, the variation does not
correlate with R, values. Furthermore, fusions expressed at
identical levels (ME10, MElO/PC599-683, and MElO/PC685815; Figure 4) demonstrate that, at least in this one case,
phytochrome sequences are mediating dimerization. The second region that we mapped, N1049-Ql129, never accumulated
to high enough levels to reliably quantitate by immunoblotting.
Therefore, an in vitro DNA binding assay was used to demonstrate that while the MElO/PC914-1129 and ME10 protein
concentrations varied with IPTG concentration in nearly identical fashion, the clone containing phytochrome sequences
was better able to repress lacZ synthesis.
The possible structure of these regions and the significance
of phytochrome dimerization are discussed below.
Structure Prediction of the Dimerization Regions
' 2; ' ;a'
tIPTG1 (WW
I
i5'
' A' ' $0
Figure 5. Levels of Active MElO Fusion Protein Assayed by Filter
Binding.
Cells containing either pMElO (filled symbols) or pMElO/PC914-1129
(opensymbols) were grown in increasingamounts of IPTG.Crudecellular lysates were preparedby freeze-thaw extraction and assayed for
total amount of active repressor by filter binding in the presence of
excess labeledoperator DNA (squares). PGalactosidaselevels for each
extract were also determinedby quantitative immunoblotting(circles).
Values shown are from a single experiment. Repeat experimentsyielded
qualitatively similar results, but differed in absolute levels of operator
bound (OR1),perhaps due to differences in extraction efficiency.
A comparison of the oat phytochromedimerization sequences
to PhyA protein sequences from five other plants in Figure 6
shows that the regions between amino acids V623-SW3 and
N1049-Ql129 are well conserved. The sequences between
V623 and S673 are particularly well conserved, with 72% amino
acid identity between corn and pea, the least similar of the
six PhyA proteins examined. In comparison,the entire carboxyterminal domains have only 58% amino acid identity between
corn and pea. This indicatesthat these conserved regions may
be important in phytochrome function.
The region between V623 and S673 contains three stretches
that could potentially form amphipathic a helices, shown as
166
The Plant Cell
AA#
helix c
helix a
AA#
helix b
AA#
AEL M A ~ ~ F E E D NEK
ADL L S ~ Y E D D N KE
PTAMGQ
PAK...
PSAVGR
NK.
HKSRTT
HKLKG.
...
1129
1128
1131
1122
1124
1125
Figure 6. Conservation of Phytochrome Dimerization Regions.
Sequence for the dimerization regions between V623-S673, N1049-Q1129,and flanking sequences of six different PhyA proteins (Hershey et al.,
1985; Sharrock et al., 1986; Sato, 1988; Christensen and Quail, 1989; Kay et al., 1989; Sharrock and Quail, 1989). Putative helices a, b, c, and
d are boxed. Residues that are 100°!o identical between all six phytochrome sequences are highlighted. Gaps in the sequence are represented
by dots. AA, amino acid; A.t., Arabidopsis.
helical wheel plots in Figure 7. Although we recognizethe limitations of secondary structuralpredictions(e.g., Nishikawaand
Noguchi, 1991),for brevity we will refer to these potential helices
as helices throughout this communication. The helices have
charged and hydrophobic residues alternating with a periodicity of three to four residues. Chou-Fassman and
Garnier-Osguthorpe-Robosonpredictionsof secondary structure (Devereux et al., 1984) assign helix propensity for the
region between V623 and S673, particularly helix c (Figure 7;
predictions not shown). We have defined the ends of these
helices by choosing the preferred N- and C-cap residues
described by Richardson and Richardson (1988). The amphipathic character of these three helices is conserved in all
of the PhyA protein sequences examined. Helices a and c (Figure 7) are also well conserved between protein sequences
encoded by two other membersof the phytochromegene family, phyS and phyC.
Amphipathic a helices are often involved in protein-protein
interaction and both heterotypic and homotypic dimerization
(Segrest et al., 1990). The free energy change upon dimerization can be estimated from the energy of water exclusion for
each helix (Richards, 1977). Using this method, the free energy
of dimerization is calculated as 8.85, 8.7, and 4.8 kcallmol for
helices a, b, and c (Figure7). As a comparison, the free energy
of dimerization for the 24-amino-acid-long helix c of catabolite gene activation protein is 7.6 kcallmol when calculated by
this method (McKay et al., 1982).
Two other points of interest on the structure of the region
between V623 and S673 are that (1) it is likely to be buriedwithin
the protein and (2) it is adjacent to areas that are probably
differentially exposed to solvent in Pr and Pfr forms. Computer
predictions of surface probability suggest that the region between roughly amino acids S620 and E690 are located within
the interior of phytochrome. More convincingly, this region is
not readily accessible to proteases (Grimm et al., 1988), and
a monoclonal antibody, GO-1, specific to this region is only
able to recognize phytochrome by immunoblotting, and not
by ELISA (Pratt et al., 1991). Sequences on either side of V623S673 are modified by a variety of agents when phytochrome
is in the Pfr conformation. For example, the regions between
L559-V669 and V748-G831 are thought to be ubiquitinatedas
Pfr (Shanklin et al., 1989), K753 is more accessible to proteases
as Pfr than Pr (Grimm et al., 1988), and S599 is preferentially
phosphorylated in the Pfr conformation (McMichael and
Lagarias, 1990).
Sequences between N1049 and Q1129 contain a second region identifiedas being involved in protein-protein interaction.
Overall, this segment is no better conserved than the entire
carboxy-terminal domain. There is 58% amino acid identity
between pea and corn carboxy-terminaldomains, versus 56%
for this region. However,sequences between N1094 and G1110
show 71% amino acid identity between pea and corn. These
sequences also have the potential to form an amphipathic a
helix, depicted in a helical wheel plot in Figure 8. As shown
in Figure 8, helix d, like helices a and c (Figure 7), is well conserved between proteins encoded by phyA, phyS, and phyC.
However, PhyC protein contains an additional residue in what
would be the center of helix d (Figure 8). Helix d, defined using
preferred N- and C-cap residues, has a predicted free energy
of water exclusion of 8.15 kcal/mol. An analysis of helical hydrophobic moments for four phytochrome proteins has also
suggested that the region around helix d (Partis and Grimm,
Phytochrome Oimerization
M622
A629
phyA con: S E M V R L y E T A T V P
A.t. phyB: R E M V R L I E T A T V P
A.t.phyC: N E M V R L I D T A A V P
167
1990) as well as the regions around helices a and c have the
potential to form an amphipathic a helix (Parker and Song,
1990).
The leucine zipper is a common dimerization motif involving amphipathic a helices. Attempts to model the phytochrome
dimerization sequences as coiled coils using both the algorithm
of Lupas et al. (1991) and molecular dynamics simulations
based on the leucine zipper structure described by OShea
et al. (1991) have failed to support a coiled coil structure for
these regions.
Jones and Quail (1986) found that an 80-kD tryptic
phytochrome fragment containing an intact amino terminus
appeared to behave in SEC more like a monomer, implying
that the dimerization region should lie in the carboxy-terminal
40-kD region (approximately residues 750 to 1129). Keeping
in mind that SEC has low resolution, the data of Jones and
Quail(l986) taken together with this work suggest that the region identified between V623 and S673, although important
in protein-protein interactions, may not be sufficient to maintain phytochrome as a dimer.
Significance of the Dimerization Regions
"642
A649
phyAcon: G L V N G W N
7K i A E LTGL
A.t. phyB: G C I N G W N A K I A E L T G L
A.t. phyC: G V I N G W N S K A A E V T G L
Severa1 plant developmental responses are regulated by
phytochrome over a wide fluence range. The dimeric model
of phytochrome action proposed by VanDerWoude (1985)
postulates that at very low fluence levels, 0.1 to 100 nmolh"
(e.g., Jones et al., 1991), plants have the ability to detect mixed
phytochrome dimers, i.e., dimers in which one member of a
pair is in the Pr conformation and the second member in the
Pfr conformation. At higher fluences, 0.01 to 1.0 mmol/m2,
plants measure the total amount of Pfr-Pfr dimers present.
Thus, in VanDeWoude's model, dimerization extends the range
of fluence sensitivity by three or more orders of magnitude.
Although VanDerWoude's model is generally accepted, it
should be noted that an alternative model by Blaauw-Jansen
(1983) has been proposed that is not based upon the dimeric
nature of phytochrome. Blaauw-Jansen proposes a fluencesensitive activation and deactivation of specific Pfr proteases.
Figure 7. Consensus Sequence and Helical Wheel Plots of Putative
Amphipathic a Helices between S620 and S673.
phyAcon: e A I G E H i L T L V E D
D
A.t. phyB: E A M G K S L V S v L I Y K
A.t. phyC: Q A I G K P * V S A L V E D
D
Helices c, a, and b are plotted in a helical wheel format to demonstrate the amphipathic character. The sequence shown is from the oat
phyA gene product. Hydrophobic amino acids are representedby stip
pled circles. Charged amino acids are indicated by (+) and (-) signs.
The consensus (con) sequence based upon the six phytochrome sequences shown in Figure 6 is shown below each wheel. Upper case
letters represent five or six identical residues; lower case letters represent three or four identical residues; two residues sharing the same
line indicate positions at which all monocots had one residue (top)
and all dicots a second residue (bottom). The corresponding sequences
of the Arabidopsis (A.t.) phyB and phyC gene products are shown beneath each consensus sequence. Dots represent gaps, and carets
represent the position of a residue.
168
The Plant Cell
METHODS
Bacterial Strains and Cloning
1109
D1102
L1095
1106
7
phyAcon: g L 1 L M N G D V R 1 R E A G
A.t.phyB:K I L K L M N G E V Q Y I R E S E
A.t.phyC: K L V K L M E,G T L R Y L R E S E
R
Figure 8. Consensus Sequence and Helical Wheel Plot of Putative
Amphipathic a Helix between N1049 and Q1129.
Helix d is plotted in helical wheel format to illustrate the amphipathic
character. The consensus sequence and the corresponding sequences
of the phy6 and phyC gene products are shown. Symbol usage is as
in Figure 7.
Physiologicaldata (e.g., Takimoto and Saji, 1984), biochemical data (Pratt et al., 1991), and DNAsequence data (Sharrock
and Quail, 1989) have shown that at least two different pools
of phytochromeexist within plant cells. Pratt et al. (1991) have
shown that heterodimers form between the phyA gene product and at least one of these other types of phytochrome,
possibly phyB andlor phyC gene products. The mapping of
potentialdimerizationsites within PhyA proteins allows a comparison of these protein sequences with similar sequences
encoded by the other members of the phytochrome gene family.
We have shown that these regions contain sequences with
the potential to form amphipathic a helices. The helices are
conserved among members of the PhyA protein family and
between PhyA proteins and Arabidopsis PhyB and PhyC proteins. The formation of homodimersand heterodimers has been
shown to play an important role in modifying the activity of
severa1 regulatory proteins such as calmodulin (ONeill and
DeGrado, 1990), NF-~6/p50(Urban et al., 1991), and cAMPdependent protein kinase (Carr et al., 1991). It is possible that
the fluence range, magnitude, andlor nature of phytochrome
responses are modulated through the formation of heterodimers comprised of PhyA protein and other types of
phytochromes.
The plasmid pMElO was constructed by inserting aT4 DNA polymerasetreated 272-bp EcoRI-Sphl fragment containing the lacUV5 promoter
and sequences encoding h repressor amino acids 1 to92 from KHlOl
(Hu et al., 1990) into EcoRV-digestedpSP72 (Promega, Madison, WI)
using standard cloning techniques (Sambrook et al., 1989). Fusion
proteins were generated by inserting fragments from pCIB315
(Thompson et al., 1989), which contains sequences from oat phyA AP3,
into pME10. MElOlGCN4 was assembled by inserting an Xbal-EcoRI
fragment from LexA/GCN4 (Hope and Struhl, 1987) encoding GCN4
residues L169-R281 into the Smal site of pMElO after the fragment
was blunted by treatment with the Klenow fragment of DNA polymerase I.All fusions were verified by restriction mapping and immunoblot
analysis using antirepressor and/or antiphytochrome antibodies. For
immunoblot analysis of fusion protein size, proteins were expressed
in insoluble form in XL-1 Blue (Stratagene, La Jolla, CA).
Repressor activity was assayed in Escherichia coli X90K (ara
Alacpro73 nalA arg€[am] rif thi-7 [F'lac pro laclq lacZ::kan]) lysogenic
for h200. A200 contains a lacZ fusion under the control of the h rightward promoter, PR(Meyer et al., 1980).
PGalactosidase Assay and Relative Repression
p-Galactosidaseactivity was assayed as described by Miller (1972) using
SDS and chloroform cell lysis. We have defined a coefficient of relative repression (R,) as being equal to the IPTG concentration at which
a given repressor has half-maximalactivity divided by the IPTG concentration at which the protein encoded by pMElO has half maximal
activity. This value was found to be a more reliable indicator of repressor function than p-galactosidase values taken at a single IPTG
concentration. The curves of P-galactosidase activity dependence on
IPTG induction were performed at least three times. In each experiment, pMElO expression was also measured as a control.
SDS-PAGE and lmmunoblots
Crude extracts for immunoblotting were produced by growing 50-mL
cultures in Luria broth with antibioticsto an ODWoof 0.6, adding IPTG
to 0.1 mM, and allowing growth to continue for 2 hr. The induced cells
were collected by centrifugation, washed in 10 mL of protein resuspension buffer (50 mM Tris, pH 8.0, 100 mM NaCI, 5 mM EDTA, 140 mM
2-mercaptoethanol, 10% glycerol), and resuspended in 10 mL of protein resuspensionbuffer. Cells were lysed by passage through a Carver
press. Unbroken cells and cell debris were removed by centrifugation
at 27,0009 for 20 min. Soluble proteins were concentrated by precipitation with 75% ammonium sulfate and resuspended in a final volume
of 500 pL of protein resuspensionbuffer. Protein concentrationswere
estimated using micro-Bradford assays (Bradford, 1976).
SDS-PAGE was carried out as described by Laemmli (1970) with
slight modification. Proteins were transferred to nitrocellulosefor immunoblot analysis. The filters were blocked for 1 hr in blocking buffer
(5% nonfat dry milk in 1 x TBS [I0 mM Tris, pH 7.45, 150 mM NaCI,
0.05% NaN,]) at room temperature. lncubation with primary antibody
was carried out in incubation buffer (0.1% BSA in 1 x TBS) with 10%
Phytochrome Dimerization
crude E. coliextract overnight at 4%. The crude bacterial cell extract
used in the incubation step was prepared by growing the appropriate
E. coli strain to an ODmoof 1.0 with 0.5 mM IPTG. The cells were then
harvested by centrifugation, resuspended in 1/50th volume protein
resuspension buffer, and lysed by passage through a Carver press.
The lysate was then cleared at 27,0009 for 20 min.
Anti-X repressor antiserum was generated as described by Jones
et al. (1991) using insoluble full-length 1repressor as an antigen. Antisera to phytochrome (4032 and 3032) were kindly provided by Peter
Quail (University of California at Berkeley, Berkeley, CA).
Size Exclusion Chromatography
Crude extract from cells expressing the pMElO/GCN4 protein was prepared from cells as described above. The extract was further purified
by the addition of 10% polyethyleneimine to a final concentration of
0.2% followed by centrifugation at 27,0009 for 20 min. Saturated ammonium sulfate was added to the supernatant to 75% saturation. Protein
was precipitated by centrifugation at 27,000s for 20 min and resuspended in 200 WL of protein resuspension buffer. Fusion protein
concentration was estimated by comparison of extract stained with
Coomassie Brilliant Blue R 250 to known concentrations of carbonic
anhydrase after separation by SDS-PAGE. Approximately 6.5 mg of
crude extract was separated over a 0.75 x 30 cm Spherogel TSK
2000SW size exclusion column (Beckman, Berkeley, CA) at O5 mUmin
in TSK buffer (50 mM Tris, pH 7.0,50 mM NaCI, 0.5 mM EDTA, 5 mM
2-mercaptoethanol, 10% glycerol). Fractions (0.25 mL) were collected
and analyzed by SDS-PAGE, followed by immunoblot analysis. A rough
estimate of a dimerizationconstant for the MElO/GCN4 fusion protein
shown in Figure 2 was determined by finding a load concentrationwhere
approximately half of the fusion protein appeared dimeric. This was
not possible with the low leve1 of expression observed for the
phytochrome fusion proteins.
Filter Binding
Filter binding assays were carried out in a manner similar to that described by Johnson et al. (1980). Cells were grown in media A (Miller,
1972) to an ODmoof 0.4 with varying amounts of IPTG. Cells (1.5 mL)
were harvestedby centrifugation at 12,0009, washed once with 500 pL
of protein resuspension buffer, repelleted, and resuspended in 30 pL
of protein resuspension buffer. The cells were then frozen in liquid
nitrogen and thawed on ice. The freeze-thaw cycle was repeated, and
the lysate was centrifuged for 20 min at 28,0009. This supernatant was
used for filter binding assays and also assayed for P-galactosidase
content by quantitative immunoblot analysis.
Binding reactions were carried out in 50 pL of 10 mM Pipes, pH
6.6, 50 mM KCI, 2 mM CaCI,, 0.1 mM EDTA, 100 pg/mL BSA, 5%
DMSO to which had been added 8.5 pg of sheared calf thymus DNA,
and 50 pmol of 32P-labeleddouble-stranded oligonucleotides containina an OR,
... bindina site (5’-GATCCTATCACCGCCAGAGGTAGGATC-3!
specific activity 500 cpmlpmol). Ten microlitersof freeze thaw supernatant was assayed for repressor content. The reaction was incubated
on ice for 1 hr, followed by filtration through nitrocellulose (0.45 pm,
NitroME; Micron Separations Inc., Westboro, MA) in a slot blotting system (Minifold (I;Schleicher and Schuell, Keene, NH). Each reaction
was washed three times with 0.3 mL of wash buffer (10 mM Tris, pH
-
I
169
7.0, 50 mM KCI, 2 mM CaCI2, 0.1 mM EDTA). The filter was then allowed to dry, and individual slots were cut out and counted.
ACKNOWLEDGMENTS
We would like to thank Dr. Jim Hu, Massachusetts lnstitute of Technology, for the h repressor clones and h200 and 1202, Dr. Peter Quail,
University of Californiaat Berkeley/U.S. Department of Agriculture-Plant
Gene Expression Center, for antibodies to phytochrome, Dr. Alex
Tropsha, University of North Carolina, for help in molecular modeling
of the dimerization region, and Drs. Mary-Michelle Cordonnier-Pratt
and Lyle Crossland, CIBA-GEIGY, for phytochrome clone pCIB315. We
would also like to express our sincere thanks to Susan Whitfield and
Deborah Cochran for their technical assistance. This work was supported by a grant to A.M.J. from the U.S. Department of Agriculture
Competitive Research Grants Office.
Received November 8, 1991; accepted January 6, 1992.
REFERENCES
Blaauw-Jansen, G. (1983). Thoughts on the possible role of
phytochromedestruction in phytochromecontrolledresponses. Plant
Cell Environ. 6, 173-179.
Bradford, M.M. (1976). A rapid and sensitive method for quantitation
of microgram quantities of protein utilizing the principle of proteindye binding. Anal. Biochem. 72, 248-254.
Carr, D.W., Stofko-Hahn, R I . , Fraser, I.D.C., Bishop, S.M., Acott,
T.S., Brennan, R.G., and Scott, J.D. (1991). lnteractionof the regulatory subunit (RII) of cAMP-dependent protein kinase with RI1
anchoring protein occurs through an amphipathic helix binding motif.
J. Biol. Chem. 266, 14188-14192.
Choi, J.-K., Kim, I.&, Kwon, T A , Parker, W., and Song, P.6 (1990).
Spectral perturbationsand oligomer/dimerformation in 124-kilodalton
Avena phytochrome. Biochemistry 29, 6883-6891.
Christensen, A.H., and Quail, P.H. (1989). Structure and expression
of a maize phytochrome-encoding gene. Gene 85, 381-390.
Cordonnier, M.-M. (1989). Monoclonal antibodies: Molecular probes
for the study of phytochrome. Photochem. Photobiol. 49, 821-831.
Devereux, J., Haeberli, P., and Smithies, O. (1984). A comprehensive set of sequence analysis programs for the VAX. Nucl. Acids
Res. 12, 387-395.
Grimm, R., Eckerskorn, C., Lottspeich, F., Zenger, C., and Rudiger,
W. (1988). Sequence analysis of proteolytic fragments of 124kilodalton phytochrome from etiolated Avena sativa L.: Conclusion
on the conformation of the native protein. Planta 174, 396-401.
.
Hecht, M.H.9 Nelson, H.C.M.9 and Sauer, R.T. (1983). Mutations in
5 repressor’s amino-terminal domain: lmplications for protein stability and DNA binding. Proc. Natl. Acad. Sci. USA 80, 2676-2680.
Hershey, H.P., Barker, R.F., Idler, K.B., Lissemore, J.L., and Quail,
P.H. (1985). Analysis of cloned cDNA and genomic sequences for
phytochrome: Complete amino acid sequences for two clonsd
170
The Plant Cell
gene products expressed in etiolated Avena. Nucl. Acids Res. 13,
8543-8559.
Hope, I.A., and Struhl, K. (1987). GCN4, a eukaryotic transcriptional
activation protein, binds as a dimer to target DNA. EM60 J. 6,
2781-2784.
Hu, J.C., OShea, E.K., Kim, P.S., and Sauer, R.T. (1990). Sequence
requirements for coiled-coil interactions: Genetic analysis using 1
repressor-GCN4 leucine zipper fusions. Science 250, 1400-1403.
Johnson, A.D., Pabo, C.O., and Sauer, R.T. (1980). Bacteriophage
h repressor and cro protein: lnteractionswith operator DNA. Methods
Enzymol. 65, 839-856.
Jones, A.M., and Erickson, H.P. (1989). Domain structure of
phytochrome from Avena sativa visualized by electron microscopy.
Photochem. Photobiol. 49, 479-483.
Jones, A.M., and Quail, P.H. (1986). Quaternary structure of 124kilodalton phytochrome from Avena sativa L. Biochemistry 25,
2987-2995.
Jones, A.M., Cochran, D.S., Lamerson, P.M., Evans, M.L., and
Cohen, J.D. (1991). Red light-regulated growth. 1. Changes in the
abundance of indoleacetic acid and a 22-kilodaltonauxin-binding
protein in maize mesocotyl. Plant Physiol. 97, 352-358.
Jones, A.M., Vierstra, R.D., Daniels, S.M., and Quail, P. (1985). The
role of separate molecular domains in the structureof phytochrome
from etiolated Avena sativa L. Planta 164, 501-506.
Kay, S.A., Keith, B., and Chua, N.-H. (1989). The sequence of the
rice phytochrome gene. Nucl. Acids Res. 17, 2865-2866.
Parker, W., and Song, P.4. (1990). Location of helical regions in
tetrapyrrole-containingproteins by a helical hydrophobic moment
analysis. J. Biol. Chem. 265, 17,568-17,575.
Partis, M.D., and Grimm, R. (1990). Computer analysis of phytochrome
sequences from five species: lmplicationsfor the mechanism of action. 2. Naturforsch. 45c, 987-998.
Pratt, L.H., Stewart, S.J., Shimazaki, Y., Wang, Y.-C., and
Cordonnier, M.-M. (1991). Monoclonal antibodies directed to
phytochromefrom green leaves of Avena sativa L. cross-react weakly
or not at all with the phytochrome that is most abundant in etiolated
shoots of the same species. Planta 184, 87-95.
Richards, F.M. (1977). Areas, volumes, packing and protein structure.
Annu. Rev. Biophys. Bioeng. 6, 151-176.
Richardson, J.S., and Richardson, D.C. (1988). Amino acid preferences for specific locations at the ends of a helices. Science 240,
1648-1652.
Riggs, A.D., Suzuki, H., and Bourgeois, S. (1970). lac repressor-operator interaction. I. Equilibrium studies. J. MOI. Biol. 48, 67-83.
Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular Cloning. A Laboratory Manual. (Cold Spring Harbor, N Y Cold Spring
Harbor Laboratory).
Sato, N. (1988). Nucleotide sequence and expression of the
phytochrome gene in Pisum sativum: Differential regulation by light
of multiple transcripts. Plant MOI.Biol. 11, 697-710.
Sauer, R.T., Jordan, S.R., and Pabo, C.O. (1990). 1 repressor: A model
system for understanding protein-DNA interactions and protein stability. Adv. Protein Chem. 40, 1-61.
Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of basteriophage T4. Nature 227, 680-685.
Segrest, J.P., De Loof, H., Dohlman, J.G., Brouillette, C.G., and
Anatharamiah, G.M. (1990). Amphipathic helix motif: Classes and
properties. Protein Struct. Funct. Genet. 8, 103-117.
Lagarias, J.C., and Mercurio, F.M, (1985). Structure function studies
on phytochrome. ldentification of light-induced conformational
changes in 124-kDaAvenaphytochrome in vitro. J. Biol. Chem. 260,
2415-2423.
Shanklin, J., Jabben, M., and Vierstra, R.D. (1989). Partia1purification and peptide mapping of ubiquitin-phytochromeconjugates from
oat. Biochemistry 28, 6028-6034.
Lupas, A., Van Dyke, M., and Stock, J. (1991). Predictingcoiled coils
from protein sequences. Science 252, 1162-1164.
McKay, D.B., Weber, I.T., and Steitz, TA. (1982). Structure of catabolite
gene activator protein at 2.9 %, resolution. J. Biol. Chem. 257,
9518-9524.
McMichael, R.W., and Lagarias, J.C. (1990). Phosphopeptide mapping of Avena phytochrome phosphorylation by protein kinases in
vitro. Biochemistry 29, 3872-3878.
Meyer, B.J., Mauer, R., and Ptashne, M. (1980). Gene regulation at
the right operator (OR) of bacteriophage 1. J. MOI. Biol. 139,
163-194.
Miller, J. H. (1972). Experimentsin Molecular Genetics. (Cold Spring
Harbor, NY Cold Spring Harbor Laboratory).
Nishikawa, K., and Noguchi, T. (1991). Predicting protein secondary
structure'based on amino acid sequence. Methods Enzymol. 202,
31-44.
ONeill, K.T., and DeGrado, W.F. (1990). How calmodulin binds its
targets: Sequence independent reccgnition of amphipathic a helices.
Trends Biol. Sci. 15, 59-64.
OShea, E.K., Klemm, J.D., Kim, P.S., and Alber, T. (1991). X-raystructure of the GCN4 leucine zipper, a two-stranded, parallel coiled coil.
Science 254, 539-544.
Sharrock, R.A., and Quail, P.H. (1989). Nove1 phytochrome sequences
in Arabidopsis thaliana:Structure, evolution, and differential expression of a plant regulatory photoreceptor family. Genes Dev. 3,
1745-1757.
Sharrock, R.A., Lissemore, J.L., and Quail, P.H. (1986). Nucleotide
and amino acid sequence of a Cucurbita phytochrome cDNA clone:
ldentification of conserved features by comparison with Avena
phytochrome. Gene 47, 287-295.
Smith, H. (1982). Light quality, photoperception, and plant strategy.
Annu. Rev. Plant Physiol. 33, 481-518.
Takimoto, N., and Saji, H. (1984). A role of phytochrome in photoperiodic induction: Two-phytochrome-pool theory. Physiol. Plant. 61,
675-682.
Thompson, L.K., Pratt, L.H., Cordonnier, M.-M., Kadwell, S., Darlix,
J.-L., and Crossland, L. (1989). Fusion protein-based epitope mapping of phytochrome. J. Biol. Chem. 264, 12426-12431.
Urban, M.B., Schneck, R., and Baeuerle, P.A. (1991). NF-KB contacts DNA by a heterodimer of the p50 and p65 subunit. EMBO J.
10, 1817-1825.
VanDerWoude, W.J. (1985). A dimeric mechanism for the action of
phytochrome: Evidence from photothermal interactions in lettuce
seed germination. Photochem. Photobiol. 42, 655-661.
VanDerWoude, W.J. (1987). Application of the dimeric model of
phytochrome action to high irradiance responses. In Phytochrome
Phytochrome Dimerization
and Photoregulationin Plants, M. Furuya, ed (Orlando, FL: Academic
Press, Inc.), pp. 249-258.
Welss, M.A., Pabo, C.O., Karplus, M., and Sauer, R.T. (1987). Dimerization of the operator binding domain of phage h repressor.
Biochemistry 26, 897-904.
171
Yamamoto, K.T., and Tokutomi, S.(1989). Formation of aggregates
of tryptic fragments derived from the carboxyl-terminal half of pea
phytochrome and localizationof the site of contact betweenthe frag
ments by amino-terminalamino acid sequence analysis. Photochem.
Photobiol. 50, 113-120.
Localization of protein-protein interactions between subunits of phytochrome.
M D Edgerton and A M Jones
Plant Cell 1992;4;161-171
DOI 10.1105/tpc.4.2.161
This information is current as of June 17, 2017
Permissions
https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw15322
98X
eTOCs
Sign up for eTOCs at:
http://www.plantcell.org/cgi/alerts/ctmain
CiteTrack Alerts
Sign up for CiteTrack Alerts at:
http://www.plantcell.org/cgi/alerts/ctmain
Subscription Information
Subscription Information for The Plant Cell and Plant Physiology is available at:
http://www.aspb.org/publications/subscriptions.cfm
© American Society of Plant Biologists
ADVANCING THE SCIENCE OF PLANT BIOLOGY