(Received for publication, August 7, 1996, and in revised form, October 4, 1996)
From the Department of Microbiology and Molecular Genetics, Molecular Biology Institute, UCLA, Los Angeles, California 90095-1489
Protein farnesyltransferase (FTase), a
heterodimer enzyme consisting of and
subunits, catalyzes the
addition of farnesyl groups to the C termini of proteins such as Ras.
In this paper, we report that the protein substrate specificity of
yeast FTase can be switched to that of a closely related enzyme,
geranylgeranyltransferase type I (GGTase I) by a single amino acid
change at one of the three residues: Ser-159, Tyr-362, or Tyr-366 of
its
-subunit, Dpr1. All three Dpr1 mutants can function as either
FTase or GGTase I
subunit in vivo, although some
differences in efficiency were observed. These results point to the
importance of two distinct regions (one at 159 and the other at 362 and
366) of Dpr1 for the recognition of the protein substrate. Analysis of
the protein, after site directed mutagenesis was used to change Ser-159
to all possible amino acids, showed that either asparagine or aspartic acid at this position allowed FTase
to function as GGTase I
. A
similar site-directed mutagenesis study on Tyr-362 showed that leucine,
methionine, or isoleucine at this position also resulted in the ability
of mutant FTase
to function as GGTase I
. Interestingly, in both
position 159 and 362 substitutions, amino acids that could change the
protein substrate specificity had similar van der Waals volumes.
Biochemical characterization of the S159N and Y362L mutant proteins
showed that their kcat/Km values for GGTase I substrate are increased about 20-fold compared with
that of the wild type protein. These results demonstrate that the
conversion of the protein substrate specificity of FTase to that of
GGTase I can be accomplished by introducing a distinct size amino acid
at either of the two residues, 159 and 362.
Protein farnesyltransferase (FTase)1
catalyzes the transfer of a farnesyl group from farnesyl diphosphate
(FPP) to a cysteine residue of a protein substrate such as Ras
(Refs. 1 and 2; for a review, see Refs. 3, 4, 5, 6). The enzyme recognizes a
tetrapeptide sequence called the CAAX motif (C is cysteine, A is an aliphatic amino acid, and X is the
C-terminal amino acid, which is usually methionine, serine, alanine,
cysteine, or glutamine) at the C terminus of the protein. A closely
related enzyme, geranylgeranyltransferase type I (GGTase I), recognizes
the CAAX motif when X is either a leucine or
phenylalanine (CAAL motif), but transfers a geranylgeranyl group from geranylgeranyl diphosphate (GGPP) (7, 8). The CAAL motif is found at the C termini of a variety of Rho
family G-proteins (3, 4, 5, 6). Although FTase and GGTase I appear to have
distinct functions inside the cell, the enzymes are not absolutely
specific in vitro; purified FTase could utilize GGTase I
protein substrates, and GGTase I could utilize FTase substrates, albeit
at a lower rate (9, 10, 11, 12). In addition, overproduction of GGTase I
-subunit could suppress temperature-sensitive growth due to the
mutation of FTase
(13, 14).
The property of FTase to recognize the CAAX motif was recently exploited to develop inhibitors of FTase. Farnesylation of Ras is critical for the membrane localization and the transforming activity of Ras. Thus, FTase inhibitors could act to suppress ras-dependent transformation (15, 16). Peptidomimetics were developed by modifying CAAX tetrapeptides, and shown both to inhibit the growth of ras-transformed cells in soft agar and to induce morphological reversion of the transformed cells (17, 18). They also inhibited the growth of ras-induced tumors in mice (reviewed in Ref. 19). Understanding the mechanism by which CAAX and CAAL substrates are recognized by the enzymes responsible for prenyl modification, and the efficiency to which this is achieved provides valuable clues to how the enzyme action can be modulated at the level of protein substrate recognition.
In order to investigate the molecular basis for the recognition of the
CAAX motif by FTase, we sought to convert protein substrate specificity of FTase to that of GGTase I. This was accomplished by
randomly mutating the -subunit of FTase. FTase and GGTase I are
heterodimers consisting of a common
subunit and a similar but
distinct
subunit (3, 20). Cross-linking of FTase with Ras or with
peptide substrates implicated the
subunits as the subunit
specifying protein substrate specificity of these enzymes (21, 22). We
have employed the yeast Saccharomyces cerevisiae as a model
system, since yeast FTase and GGTase I, like mammalian enzymes, are
composed of a common
subunit encoded by the RAM2 gene
and related
subunits encoded by the DPR1/RAM1
gene and the CAL1/CDC43 gene, respectively
(23, 24, 25, 26, 27). Dpr1 protein shares significant sequence identity with the
human FTase
subunit and with Cal1 protein (14, 28). In addition, yeast FTase shows a divalent cation requirement and a protein substrate
specificity similar to those of mammalian enzyme (12, 29). We have
developed a screen to identify mutant FTase
, which could function
as a GGTase I
(30). The mutant FTases enable farnesylation of
GGTase I substrates, which is sufficient to support GGTase I function
inside the cell.
Previously, we reported the initial isolation of a mutant FTase subunit, S159N, which replaced the function of GGTase I
(30). In
the current study, we carried out an extensive screen to identify
mutants with altered specificity for protein substrates similar to that
reported for S159N. Two mutants in the C-terminal portion of the Dpr1
molecule (Y362H and Y366N) were identified in addition to S159N.
Site-directed mutagenesis at the position 159 and 362 residues was
carried out to gain further insights into the molecular mechanism by
which these residues affect substrate specificity. Our results suggest
that introduction of an amino acid of a specific size at residue 159 or
at residue 362 results in the conversion of the protein substrate
specificity of FTase to that of GGTase I.
S.
cerevisiae strains used were YOT559-3C (MATa
cal1-1 leu2 trp1 ura3 ade2) (14) and YPH250dU
(MATa dpr1::URA3 lys2-801
leu2-1 trp1-
1 ura3-52 ade2-101
his3-
200). YPH250dU was constructed as follows. The
2.4-kb dpr1::URA3 fragment was amplified by PCR
from the genomic DNA of strain KMY200-sgp2-12A (31) using the primers
DPR1#1 and DPR1#2 (30) and was used to transform strain YPH250 (32) to
Ura+. The disruption was confirmed by PCR. Yeast media used
were YPD medium (33) and SC-trp medium, synthetic minimal medium (33) supplemented with adenine, uracil, and 0.5% casamino acids. Yeast transformation was carried out by the lithium acetate method (34).
Yeast high copy (YEp) plasmids containing wild type or mutant DPR1 gene were derivatives of pYO324 (35). Yeast low copy (YCp) plasmids containing wild type or mutant DPR1 gene were constructed by transferring the 2.4-kb XhoI-SacI DPR1 fragment from the high copy plasmids to the XhoI/SacI site of pRS314 (32). The DPR1 gene was randomly mutagenized by PCR with Taq DNA polymerase (Perkin-Elmer) to construct libraries on yeast high copy plasmid. To efficiently identify independent mutations, the procedure described previously (30) was modified as follows. First, the DPR1 gene was divided into two parts and was amplified separately; the 0.9-kb BamHI-NcoI region and the 0.5-kb NcoI-AflII region of pYO324(DPR1)BA were amplified using the primers DPR1#1 and DPR1#4 and the primers DPR1#2 and DPR1#3, respectively. The two regions correspond to amino acids 1-274 and 274-425, and all but the C-terminal six residues were mutagenized during the PCR amplification by inherent infidelity of Taq DNA polymerase. Second, a lower concentration of dNTP (50 µM each instead of 200 µM each) and fewer cycles of reaction (20 cycles instead of 30 cycles) were used to reduce multiple mutations, which we had found in the previous screen. Third, for each region, six libraries were constructed from independent PCR reactions to isolate independent mutations. Each library was derived from 36,000-86,000 Escherichia coli transformants. This approach is largely confined to detect single nucleotide changes in the DPR1 gene.
Isolation of Mutant DPR1 Genes That Suppress cal1A
temperature-sensitive cal1-1 strain, YOT559-3C, was used to
screen the mutated DPR1 gene libraries for suppression. This strain cannot grow at the restrictive temperature and grows slower than
wild type strain even at the permissive temperature. These phenotypes
can be suppressed by the expression of a mutant DPR1, which
can function as CAL1. In contrast, the wild type
DPR1 gene on a high copy plasmid not only fails to suppress
the temperature-sensitive growth, it also impairs the slow growth of
the cal1-1 strain even at the permissive temperature (30).
This latter property is caused presumably by the competition between
Cal1-1 protein and overproduced Dpr1 protein for Ram2, the subunit
shared by FTase and GGTase I. Thus, screening at the restrictive and at
the permissive temperatures provides a sensitive assay for
CAL1 function, allowing us to select fast growing colonies
from a plate containing several thousand colonies.
Mutant DPR1 genes that can suppress the cal1-1 mutation were identified as follows. The mutated DPR1 gene libraries were used to transform the cal1-1 strain. Fast growing transformants were selected from SC-trp plates incubated at 27 °C and then screened for growth at 37 °C. Plasmid dependent suppression was confirmed by retransformation. One clone was chosen from each library to ensure independent origins of mutations and sequenced by dideoxy sequencing with Sequenase (U. S. Biochemical Corp.) using plasmid DNAs as templates.
Protein Prenyltransferase AssaysProtein prenyltransferase assays were carried out as described (30). 5 µM ZnCl2 was included in the reaction mixture as described (30). Yeast crude soluble extracts prepared as described (30) or enzymes purified from E. coli (see below) were used for assays. Glutathione S-transferase (GST) fusion proteins, GST-CIIS and GST-CIIL (26), were used as protein substrates. Prenyl substrates were [3H]FPP (22.3 Ci/mmol) and [3H]GGPP (19.3 Ci/mmol) (DuPont NEN). Incubations were carried out using 76 nM FTase at 30 °C for the periods indicated in figure legends.
Purification of Wild Type and Mutant FTasesFTases were
purified by fusing the DPR1 with glutathione
S-transferase and using glutathione beads. For this purpose,
DPR1 was cloned into pGEX vector to produce pGEX-DPR1.
Briefly, DPR1 fragment was PCR-amplified to produce a
fragment with a BamHI site at the 5 end and a
PstI site at the 3
end of the gene. The fragment was cloned
into the BamHI and PstI sites of pBluescript II
SK. The DPR1 fragment cut out with BamHI and
EcoRI was cloned into the BamHI and
EcoRI sites of pGEX-1 (36). S159N and Y362L mutant plasmids
were constructed by using PCR to introduce each mutation into
pGEX-DPR1. A plasmid, pBC-RAM2, was constructed by placing
RAM2 into pBC-KS+ (Stratagene) in two steps (37). First, a
XbaI-SphI RAM2 fragment was cloned
into XbaI-SphI sites of pGEM-3Z (Promega) to
create pGEM-RAM2. The 5
XbaI-BglII portion of
RAM2 was replaced with the PCR product amplified from
the original RAM2 clone (24) using the primers
5
-GCTCTAGACTCGAGGAAAATGGAGGAG-3
and 5
-CGCGCGGAAGATCTGATAGTTCTTTG3
.
The XhoI-HindIII fragment of RAM2 from
this construct was cloned into the same sites of pBC-KS+ to yield
pBC-RAM2. The constructs utilizing PCR products were sequenced to
confirm the mutations and to make sure that no additional mutations
were introduced.
E. coli strain DH5 carrying pGEX-DPR1 and pBC-RAM2 was
grown in LB medium containing chloramphenicol and ampicillin to an A600 of 0.6 and then induced with 1 mM isopropyl-1-thio-
-D-galactopyranoside. The induction was carried out for 12 h at room temperature. The cell pellets were resuspended in 50 mM Tris-HCl, pH 7.4, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl
fluoride, and 0.2 mg/ml lysozyme. The cells were broken by sonication,
and lysates were centrifuged at 12,000 × g for 30 min.
Glutathione beads were added, and the solution was rotated for 2 h
at 4 °C. FTase was eluted by the addition of 10 mM
glutathione and the eluate was concentrated by Centricon 50 (Amicon).
Glycerol was added to 50% and the enzymes were kept at
20 °C.
Protein concentrations were determined by Bradford assay. This
purification yielded homogeneous preparations that contained two bands,
70 and 40 kDa, corresponding to GST-Dpr1 and Ram2 subunits,
respectively, on a SDS-polyacrylamide gel. Scanning of the gel showed
that the preparations we used were greater than 90% pure.
Site-directed mutagenesis of Ser-159 and Tyr-362 was carried out by PCR using overlap extension (38, 39) with Pfu DNA polymerase (Stratagene). Mutagenesis of Ser-159 was carried out with the flanking primers DPR1#1 and DPR1#4 and the internal, complementary primers 159S and 159A. The primer 159S contains the degenerate sequence NNN, TWC, GAM, or a sequence TGG at codon 159 (N = A, C, G, or T; W = A or T; M = A or C). Amplified fragment was digested with BamHI plus NcoI and was used to replace the corresponding region of pYO324(DPR1)BA. Plasmids from E. coli transformants were sequenced to determine the residue at 159 and to confirm that no other mutations were introduced. The degenerate primer containing NNN yielded 15 amino acid residues other than Tyr, Phe, Glu, Asp, and Trp. The primer containing TWC, GAM, or TGG yielded Tyr and Phe, Glu and Asp, or Trp, respectively. Mutagenesis of Tyr-362 was carried out with the flanking primers DPR1#2 and DPR1#3 and the internal, complementary primers 362S and 362A. The primer 362S contains the degenerate sequence NNN, RAC, WKG (N = A, C, G, or T; R = A or G; W = A or T; K = G or T) or a sequence ATC. The degenerate primer NNN yielded 15 different amino acid residues other than Asn, Gln, Trp, Met, and Ile. The primer RAC yielded Asn and Gln, whereas WKG yielded Trp and Met. The primer ATC yielded Ile. Amplified fragments were digested with NcoI and AflII and were used to replace the corresponding region of pYO324(DPR1)BA. All mutations were confirmed by dideoxy sequencing (U. S. Biochemical Corp.).
Primer Sequences
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Nucleotides that are not complementary to DPR1 are underlined. 159A and 362A are complementary to 159S and 362S, respectively.
Construction of Double MutantsDPR1 genes encoding mutations at either residues 159 or 362 were randomly joined to produce the double mutant library. Mutant DNAs at residue 159 except for the strong suppressor, Asn, were mixed in equimolar amounts and digested with NcoI and AflII to produce fragments with N-terminal mutations. Equimolar amounts of mutants at residue 362, except for the strong suppressors, Leu, Met, or Ile, were digested with NcoI and AflII to produce C-terminal fragments. The resulting fragments were ligated, and the library of double mutants was transformed into the cal1-1 strain, YOT559-3C. Transformants were grown on SC-trp at 26 °C for 3 days. Approximately 103 colonies were screened. Fast growing colonies were selected and subsequently tested for growth at restrictive temperatures. The plasmids from these clones were recovered, retransformed into cal1-1 to confirm the suppression, and sequenced.
We have previously developed a screen to identify mutant
yeast FTases with increased ability to utilize GGTase I protein
substrates (30). The screen is based on a yeast mutant cal1,
which is defective in the -subunit of GGTase I. Temperature-sensitive growth of this mutant can be suppressed by
introducing mutant DPR1 genes, which could function as
GGTase I
. The mutant DPR1 enables the efficient
utilization of GGTase I protein substrates, but attaches a farnesyl
instead of a geranylgeranyl group to the substrate proteins.
Apparently, farnesylated GGTase I substrates are competent in
suppressing cal1 phenotypes. After the initial isolation of a mutant Dpr1, S159N (30), we carried out a more extensive screen in an
attempt to identify additional mutants. The PCR mutagenesis procedure
employed in the previous screen was modified to isolate independent
mutations more efficiently (see "Experimental Procedures"). Briefly, N-terminal and C-terminal portions of the DPR1 gene
were separately amplified under conditions optimized so as to generate single point mutations in the region of amplification. Two sets of
mutated DPR1 gene libraries, one having the N-terminal
portion mutated and the other having the C-terminal portion mutated,
were constructed separately on a yeast high copy plasmid. These
libraries were then transformed into the cal1-1 strain, and
transformants capable of growth at the restrictive temperature were
selected. The plasmids from such suppressors were recovered and
sequenced. Using the N-terminal mutant libraries, we found six separate
G
A transitions at nucleotide 476 (the same mutation as the
previously identified mutation; Ref. 30), which causes a Ser-159
Asn substitution (designated S159N). Using the C-terminal mutant
libraries, we identified four T
C transitions at 1084, which cause
a Tyr
His substitution at residue 362 (Y362H); and one T
A
transversion at 1096, which causes a Tyr
Asn substitution at
residue 366 (Y366N).
Fig. 1 shows the suppression of the
temperature-sensitive growth of the cal1-1 strain by the
three mutant DPR1 genes on a high copy plasmid isolated from
the screen described above. The mutant DPR1 genes suppressed
the temperature sensitivity of the cal1-1 strain to varying
degrees; S159N showed the most efficient suppression, Y366N showed
suppression comparable to S159N, and Y362H showed the least efficient
suppression. The three mutant DPR1 genes were then
transferred to a low copy plasmid and were tested for suppression of
cal1-1 and complementation of
dpr1::URA3, a disruption allele of
DPR1. The results are summarized in Table I.
For suppression of cal1-1, wild type DPR1 gene on
a low copy plasmid further impaired the slow growth of the
cal1-1 strain at permissive temperatures, as did wild type
DPR1 gene on high copy plasmid (30), while the mutant
DPR1 genes suppressed the slow growth of the strain. Y362H
appears to suppress less efficiently than S159N and Y366N, since the
doubling time of the Y362H transformant was longer than that of the
other two transformants. At restrictive temperatures, S159N showed
suppression at both 35 °C and 37.5 °C, whereas Y366N showed
suppression at 35 °C but not at 37.5 °C. Y362H showed only weak
suppression at the restrictive temperatures. The order of the
efficiency of suppression for the three mutants is consistent with the
differences observed for the high copy plasmid (Fig. 1).
|
For the complementation of dpr1 disruption mutation, the
mutant DPR1 genes, like wild type DPR1 gene,
complemented dpr1 mutation for both the slow growth at
permissive temperatures and the temperature-sensitive growth. The
mutant DPR1 genes also complemented the sterile phenotype of
the dpr1::URA3 strain (data not shown). The
suppression of cal1 and the complementation of
dpr1 suggest that S159N, Y362H, and Y366N can act as the
-subunit for both GGTase I and FTase in vivo.
As described above, the
substitution of Ser-159 of yeast FTase subunit by Asn conferred
cal1 suppression. In order to determine what other amino
acid substitutions could affect protein substrate specificity, we
substituted Ser-159 of Dpr1 with all 20 amino acids and tested each
both for the suppression of cal1-1 and for the
complementation of dpr1::URA3. The amino acid
substitutions were accomplished by using random oligonucleotides at the
serine-159 codon. The ability of the mutant Dpr1 proteins to
functionally replace the cal1-1 mutant was assessed from
growth of transformants at permissive and restrictive temperatures. The
impaired growth of the cal1-1 strain by wild type
DPR1 at permissive temperature provided a sensitive assay
for CAL1 function (see "Experimental Procedures"). Table
II summarizes the results obtained. As can be seen, Asn
is the only residue that conferred efficient suppression of both the
slow growth at permissive temperature and the temperature-sensitive growth of the cal1-1 strain. Besides Asn, only Asp conferred
suppression of the temperature-sensitive growth at restrictive
temperatures. However, this latter mutant was not as efficient as the
Asn mutant. At the permissive temperature, Cys, Thr, Val, and His
conferred the growth of transformants at a rate intermediate between
that observed with wild type DPR1 and that observed with
vector. In other words, these residues partially relieved the impaired
growth by wild type DPR1, suggesting that Cys, Thr, Val, or
His confers a weak CAL1 activity that is intermediate
between that of DPR1 and that of CAL1. This
partial relief from the impaired growth is unlikely to be due to
instability of the protein or a defect of the protein in the
interaction with the
subunit, because these mutants conferred
complementation of dpr1. The other residues showed the same
phenotypes as wild type DPR1, i.e. impaired
growth at permissive temperature and no suppression at restrictive
temperatures.
|
The above results on amino acid residues that confer CAL1 activity can best be explained by the physical size of the amino acids. Table II shows correlation of CAL1 function with the size of the residue. It is interesting that Asn and Asp confer CAL1 function, while Gln and Glu do not. This suggests that the physical size of the residue is important rather than chemical properties. In fact, these two residues, as well as Thr, which exhibit weak CAL1 activity, have similar van der Waals volumes (91-96 Å3; Table II).
The mutant protein could contain any amino acid at 159, except Phe, Tyr, Arg, or Trp, and confer complementation of dpr1 for both the slow growth at permissive temperature and the temperature-sensitive growth. Phe, Tyr, Arg, and Trp conferred complementation for the slow growth at permissive temperature but not for the temperature-sensitive growth. We, therefore, conclude that very large residues at amino acid 159 interfere with DPR1 function.
Site-directed Mutagenesis of Tyr-362We also substituted Tyr-362, one of the tyrosine residues in the C terminus of Dpr1 with all 20 possible amino acids. Results are summarized in Table III. Three amino acid substitutions, Ile, Leu, and Met, exhibited efficient suppression of the cal1 phenotype. Of the three, Leu and Met mutants suppressed better than the Ile mutant. Interestingly, these amino acid residues have the same van der Waals volume (124 Å3; Table III), suggesting that it is the size of the amino acid that is important for conferring CAL1 activity. It is worth noting that these amino acids are hydrophobic amino acids, suggesting that the apolar character of these residues may also be important for substrate recognition of CAAL motif. Weak suppression was obtained with Thr and Val mutants at restrictive temperatures. It is rather surprising that the amino acid size (93-105 Å3; Table III) of these weak suppressors is close to the size of those identified as suppressors in the Ser-159 mutagenesis described above (see "Discussion"). The original mutant, Y362H, suppressed only weakly at 35 °C. No suppression was noted at 37 °C, even when multi-copy plasmids were used (Table III). In addition to Thr and Val, weak suppression by Gly, Cys, Asp, and Asn was seen at the permissive temperature.
|
Fig. 2 compares the effect of some of the Tyr-362
substitution mutants with the S159N mutant on the doubling time of the
cal1-1 mutant. As can be seen, the Y362M mutant grew at a
rate slightly faster than that seen with the S159N transformant. Y362L
and Y362I transformants exhibited growth rates less than that seen with the S159N transformant. A weak suppressor Y362H clearly grew better than the vector control, but its growth rate was approximately 3-fold
slower than that observed with the S159N mutant.
We next tested whether these Tyr-362 substitution mutants were capable of complementing dpr1 mutation. As shown in Table III, Arg and Lys mutants were the only ones that did not complement the temperature-sensitive growth of the dpr1 mutant. We conclude that the basic property of these amino acids may interfere with the DPR1 function.
Genetic Evidence for the Cooperation between Residues 159 and 362To gain insights into possible interaction between residues 159 and 362, we attempted to obtain double mutants having alterations of both serine 159 and tyrosine 362. A library of DPR1 containing mutations at both these residues was constructed by combining the mutant DPR1 genes identified above (see "Experimental Procedures"). Strong suppressors (S159N, Y362L, Y362M, and Y362I) were excluded from the library. The library was transformed into the cal1-1 strain, and transformants capable of suppressing cal1 were identified by their growth at restrictive temperatures. This study led to the identification of a double mutant, which exhibited significant cal1 suppression. As shown in Fig. 2, the double mutant (DM) grew at a rate comparable with that of Y362L. The double mutant contained S159D and Y362N mutations. S159D was observed to be a weak suppressor at restrictive temperatures, and Y362N was shown to exhibit weak suppression only at the permissive temperature. Thus, combining weak suppressors will produce additive effects, suggesting that the two residues 159 and 362 cooperate to recognize the CAAL motif.
Characterization of Mutant FTasesSuppression of the
cal1-1 mutation by the mutant DPR1 genes
described above suggests that mutant FTases farnesylate or
geranylgeranylate GGTase I substrates in vivo. Crude soluble
extracts, prepared from cal1-1 transformants carrying wild
type or mutant DPR1 genes on a low copy plasmid, were tested
for enzymatic activity to farnesylate or geranylgeranylate GST-CIIL
protein, a preferred substrate for GGTase I (26). FTase activity from
cal1 transformants expressing the mutant subunit S159N,
Y362H, or Y366N showed an increased ability to farnesylate the GST-CIIL
protein compared with wild type FTase, but showed little activity to
geranylgeranylate the GST-CIIL protein (data not shown). These in
vitro results suggest that the mutant FTases have altered
specificity for protein substrates but not for prenyl substrates and
that the suppression is due to the ability of the mutant enzymes to
farnesylate GGTase I substrates in vivo.
To further characterize the mutant enzymes, we purified the wild type
and two mutant enzymes, S159N and Y362L. To accomplish this, we fused
the DPR1 gene with glutathione S-transferase.
This fusion construct was transformed into E. coli cells
together with a RAM2 construct. Affinity purification with
glutathione beads yielded homogeneous preparations (see "Experimental
Procedures"). Fig. 3 shows time course of the
incorporation of farnesyl or geranylgeranyl radioactivity into FTase
substrate, GST-CIIS and into GGTase I substrate, GST-CIIL. As can be
seen in Fig. 3 (left panel), the wild type FTase enzyme
incorporated farnesyl radioactivity into GST-CIIS approximately 4 times
better than that into GST-CIIL. In contrast, both mutant enzymes, S159N
and Y362L, incorporated farnesyl radioactivity into GST-CIIS and
GST-CIIL with comparable efficiency. Virtually no incorporation of
radioactive geranylgeranyl was observed with either GST-CIIS and
GST-CIIL in the case of the wild type and S159N enzymes. In the case of
Y362L mutant, low but reproducible incorporation of radioactive
geranylgeranyl was observed with the GST-CIIS substrate. Thus, both the
S159N and Y362L mutants have increased ability to farnesylate the
GGTase I substrate. Similar conclusions can be drawn from the results of experiments altering substrate protein concentration (Fig. 4). Kinetic parameters, Km and
kcat, are summarized in Table IV.
These values were obtained from Lineweaver-Burke plots of the wild type
and mutant enzymes carried out in the presence of saturating amounts of
farnesyl diphosphate. The wild type enzyme has a Km
for GST-CIIS that is approximately 6-fold lower than the
Km for GST-CIIL. kcat for
CIIS is approximately 10-fold larger than that for GST-CIIL. Because of
these differences, kcat/Km
value for GST-CIIS is approximately 50-fold higher than the value for
GST-CIIL. Thus, the wild type enzyme prefers GST-CIIS over GST-CIIL. In
contrast, the mutant S159N shows only a 1.5-fold difference in the
kcat/Km values between
GST-CIIS and GST-CIIL. In the case of the Y362L mutant, similar
kcat/Km values were obtained
for the two protein substrates. This suggests that the mutant enzymes
farnesylate GST-CIIS and GST-CIIL substrates with an equivalent
efficiency.
|
FTase recognizes the CAAX motif, whereas GGTase I
recognizes the CAAL motif (3, 4, 5, 6). This suggests that the protein substrate binding pockets of these two enzymes have similar but
distinct conformations to allow their specificity differences. In the
current report, we show that the conversion of the FTase type
recognition to the GGTase I type recognition can be accomplished by a
single amino acid change in one of three different residues. These are
Ser-159, Tyr-362, and Tyr-366 of yeast FTase subunit, Dpr1 protein.
We believe that these residues represent major contributors for the
CAAX recognition, since the number of mutations that can be
isolated from the screen is likely to be close to saturation because
independent libraries yielded the same mutations. As schematized in
Fig. 5A, these results suggest that two
regions of Dpr1 (one surrounding Ser-159 and the other encompassing
Tyr-362 and Tyr-366) are involved in the interaction with the protein
substrate. It is possible that these residues are in close proximity to
each other in a three-dimensional conformation and form a binding
pocket for the protein substrate (see below).
Site-directed mutagenesis of Ser-159 and Tyr-362 revealed that size rather than chemical properties of the residue is important for the conversion of FTase substrate binding to GGTase I substrate binding. For suppression of cal1-1, Asn is the only residue that conferred efficient suppression at residue-159. Asp, which is different from Asn in chemical properties but is similar in size, conferred weak suppression. In contrast, Gln, which is chemically similar to but is larger than Asn, showed no suppression. Furthermore, residues that conferred very weak CAL1 function, Thr, Cys, and Val, have similar size relative to Asn despite differences in chemical properties. Thus, size appears more important than chemical properties at residue 159 for function as CAL1. A similar interpretation could be made for the results obtained with the site-directed mutagenesis of Tyr-362. In this case, the three mutants that were capable of suppressing cal1 defect contained amino acids Leu, Met, or Ile, which have identical van der Waals volume. Thus, introducing a distinct size amino acid at either 159 or 362 converts the substrate specificity of FTase to that of GGTase I. Presumably, the introduction of a particular size amino acid alters the structure of the substrate binding site. It is worth noting that tyrosine (wild type residue) has a van der Waals volume considerably larger than that of leucine. This argues against the idea that the conversion of FTase to GGTase I involves the introduction of a larger size amino acid in order to open up a pocket. Rather, a precise change is required to accomplish the conversion. In addition, it is intriguing that the three amino acids (Met, Leu, and Ile) detected at residue 362 are all hydrophobic amino acids. One interpretation is that replacement of Tyr at this residue with a smaller aliphatic residue could introduce a nice apolar binding pocket for the Leu side chain of the CIIL peptide.
We have obtained genetic evidence for the cooperativity between the residues 159 and 362. A single mutant S159D or Y362N only weakly suppresses cal1 mutant. However, the double mutant S159D/Y362N exhibits efficient suppression (Fig. 2). One possibility for this additive effect of the alterations at residues 159 and 362 of FTase is that these residues are proximally located in a three-dimensional structure. This proposal is further supported by the detection of weak suppressors of the cal1 phenotype. Weak suppression was accomplished by the conversion of Tyr-362 to Val or Thr. These residues have van der Waals volumes (91-105 Å3) similar to those amino acids that have been detected at Ser-159 to suppress cal1 mutation. Thus, introduction of a certain amino acids of approximately the same size at either 159 or 362 may result in conversion of an FTase pocket to a GGTase I pocket. A possible interpretation of these results is that the residues 159 and 362 are located in the same substrate pocket.
Biochemical analyses of the mutant proteins S159N and Y362L demonstrate that these proteins have increased affinity for protein substrates with the CAAL motif. The wild type enzyme shows approximately 50-fold higher kcat/Km value for the FTase substrate compared with the GGTase I substrate. In the case of the mutant enzymes, while kcat/Km values for the FTase substrates decrease, kcat/Km values for the GGTase I substrates are now similar to those observed for the FTase substrates. These results could explain the ability of the mutants to suppress cal1 phenotypes, since FTase and GGTase I substrates are modified with equivalent efficiency. In addition, in the case of the Y362L mutant, we observed a low but significant incorporation of radioactive geranylgeranyl into GST-CIIS protein. Thus, this latter mutant can also incorporate radioactive geranylgeranyl. Further experiments with other protein substrates are needed to assess the significance of the geranylgeranylation brought about by this mutant enzyme.
Our results also show that the residues 159 and 362 can be changed to
other amino acids without losing the ability to function as a
-subunit of FTase. Of the 20 amino acids tested for Ser-159, all
except Phe, Tyr, Arg, and Trp conferred complementation of dpr1 for temperature sensitivity. These four residues, which
did not confer complementation, are chemically different but similar in
having large van der Waals volumes. The only amino acids that did not
confer complementation in the Tyr-362 replacement studies were Arg and
Lys. In this case, the introduction of basic residues appears to
interfere with the DPR1 function.
A survey of sequences surrounding the residues 159 and 362 among FTases from a variety of sources shows that many amino acids are conserved (Fig. 5B). Tyrosine 362 is also conserved in FTases. The sequences in the corresponding regions of GGTase I, however, are different from these conserved FTase sequences. In particular, regions around residue 362 are interesting. These regions are conserved among FTase and GGTase I better than the regions around Ser-159. Some residues (e.g. Gly-350, Asp-360, and His-363) are conserved among FTase and GGTase I, and others (e.g. Leu-351, Asp-353, and Cys-367) are not conserved between FTase and GGTase I (Fig. 5B). Given the distinct but closely related specificities for protein substrates of FTase and GGTase I, this region is likely to play a significant role in protein substrate recognition and/or binding.
FTase and GGTase I are zinc metalloenzymes; both require zinc for activity (7, 40) and contain a single zinc atom per molecule (41, 42). The observation that zinc is required for the transfer of a prenyl group to protein substrates suggests a role of zinc in the recognition of protein substrates. Recently, Zhang et al. (43) showed that replacing Zn2+ to Cd2+ in FTase results in the acquisition of the ability to recognize GGTase I protein substrate. This suggests that the protein substrate specificity of FTase can be altered not only by introducing mutations in the enzyme but also by metal ion replacements. However, the exact mechanisms by how these manipulations affect substrate specificity may not be related, since our assays are carried out in the presence of ZnCl2.
In conclusion, our results provide insights into the recognition of protein substrates by FTase. Successful conversion of the FTase recognition to GGTase I recognition by the introduction of a particular size amino acid points to the remarkable versatility of protein prenyltransferases for their substrate recognition. Our data on yeast FTase described here provide valuable information for the understanding of the structure and function of FTase.
We thank Drs. Peter Edwards, David Sigman, and James Bowie for critical reading of the manuscript, and Jun Urano and Dr. Patrick Poullet for valuable comments. We also thank Kenneth Esson for DNA sequencing.