These studies have produced a set of empiric rules for amino acid require-ments at distinct positions in the N-myristoylation motif as follows: a the requirement for Gly at position 2 is
Trang 1Vertical-scanning mutagenesis of amino acids in a model
N-myristoylation motif reveals the major amino-terminal
sequence requirements for protein N-myristoylation
Toshihiko Utsumi, Kengo Nakano, Takeshi Funakoshi, Yoshiyuki Kayano, Sayaka Nakao, Nagisa Sakurai, Hiroyuki Iwata1and Rumi Ishisaka
Department of Biological Chemistry and1Department of Veterinary Medicine, Faculty of Agriculture, Yamaguchi University, Yamaguchi, Japan
In order to determine the amino-terminal sequence
requirements for protein N-myristoylation, site-directed
mutagenesis of the N-terminal region was performed using
tumor necrosis factor (TNF) mutants as model substrate
proteins Subsequently, the susceptibility of these mutants to
protein N-myristoylation was evaluated by metabolic
labe-ling in an in vitro translation system using rabbit reticulocyte
lysate A TNF mutant having the sequence MGAAAAA
AAA at its N-terminus was used as the starting sequence
to identify elements critical for protein N-myristoylation
Sequential vertical-scanning mutagenesis of amino acids at a
distinct position in this model N-terminal sequence revealed
the major sequence requirements for protein
N-myristoyla-tion: the combination of amino acids at position 3 and 6
constitutes a major determinant for the susceptibility to
protein N-myristoylation When Ser was located at position
6, 11 amino acids (Gly, Ala, S er, Cys, Thr, Val, Asn, Leu, Ile,
Gln, His) were permitted at position 3 to direct efficient protein N-myristoylation In this case, the presence of Lys at position 7 was found to affect the amino acid requirement at position 3 and Lys became permitted at this position When Ser was not located at position 6, only 3 amino acids (Ala, Asn, Gln) were permitted at position 3 to direct efficient protein N-myristoylation The amino acid requirements found in this study were fully consistent with the N-terminal sequence of 78 N-myristoylated proteins in which N-myr-istoylation was experimentally verified These observations strongly indicate that the combination of amino acids at position 3, 6 and 7 is a major determinant for protein N-myristoylation
Keywords: N-myristoylation motif; N-myristoyltransferase; protein N-myristoylation; substrate specificity; vertical scanning mutagenesis
A number of eukaryotic cellular proteins are found to be
covalently modified with the 14-carbon saturated fatty
acid, myristic acid [1–5] Many of the N-myristoylated
proteins play key roles in regulating cellular structure and
function They include proteins involved in a wide variety
of cellular signal transduction pathways In general,
protein N-myristoylation is the result of cotranslational
addition of myristic acid to a Gly residue at the extreme
N-terminus after removal of the initiating Met A stable
amide bond links myristic acid irreversibly to proteins
N-Myristoylation can also occur post-translationally, as in
the case of the pro-apoptotic protein BID and cytoskeletal
actin, where proteolytic cleavage by caspase reveals a
hidden myristoylation motif [6,7] N-Myristoylation is
catalyzed by N-myristoyltransferase (NMT), a member of
the GCN5 acetyltransferase (GNAT) superfamily of proteins [8] NMT has been purified and cloned from many organisms [9–12] and its substrate specificities have been characterized In general, Ser or Thr is preferred at position 6, and the N-terminal consensus motif Me-Gly-X-X-X-Ser/Thr that directs protein N-myristoylation has been defined [13] Saccharomyces cerevisiae NMT (NMT1p) is the best studied of the known NMTs The precise substrate specificity of this enzyme has been characterized using purified enzyme and synthetic peptides derived from the N-terminal sequences of known N-myristoylated proteins [1,14,15] These studies have produced a set of empiric rules for amino acid require-ments at distinct positions in the N-myristoylation motif
as follows: (a) the requirement for Gly at position 2 is absolute; (b) charged residues, aromatics and Pro are not allowed at position 3; (c) all amino acids are allowed at positions 4 and 5; (d) Ser, Thr, Ala, Gly, Cys, or Asn are permitted at position 6; (e) all but Pro are allowed at position 7 [5] Thus, it is well accepted that in addition to Gly at position 2, the amino acids at positions 3, 6, and 7 play an important role in substrate recognition by NMT
In fact, it was demonstrated that the difference in the substrate specificity of NMT in different species depends mainly on the difference in the permitted amino acid residues in these three positions [16,17] However, the
Correspondence to T Utsumi, Department of Biological Chemistry,
Faculty of Agriculture, Yamaguchi University, Yamaguchi 753-8515,
Japan Fax: + 81 83 933 5820, Tel.: + 81 83 933 5856,
E-mail: utsumi@yamaguchi-u.ac.jp
Abbreviations: DMEM, Dulbecco modified Eagle’s medium; DPBS,
Dulbecco’s phosphate-buffered saline; NMT, N-myristoyltransferase;
TNF, tumor necrosis factor.
(Received 1 November 2003, revised 25 December 2003,
accepted 13 January 2004)
Trang 2relative roles of these residues in substrate recognition or
the relationship between the amino acids that reside at
these three positions have not been well characterized so
far
Proteins destined to become N-myristoylated begin with
the sequence Gly However, proteins having the
Met-Gly sequence at their N-terminus may also be subjected
to another cotranslational protein modification,
N-acety-lation In fact, many proteins having an N-terminal
Met-Gly sequence, such as ovalbumin [18], cytochrome c [19],
actin [20], and 20Sproteasome a3 subunit [21] have been
found to be N-acetylated N-Acetyltransferases that
cata-lyze cotranslational protein N-acetylation also have a
restricted number of substrates [22–24] However, the
differences in the N-terminal sequence requirements for
protein N-myristoylation and protein N-acetylation have
not been fully characterized yet In a previous report, we
showed that metabolic labeling in an in vitro translation
system is an effective strategy to characterize
cotransla-tional N-terminal protein modifications [25,26] As the
in vitrotranslation system using rabbit reticulocyte lysate
contains all the components involved in cotranslational
protein N-myristoylation and N-acetylation [18,20,27], the
use of this system to study cotranslational protein
N-myristoylation seems to be appropriate Using this
assay system, we demonstrated previously that the amino
acid residue at position 3 strongly affects protein
N-myristoylation, and the amino acid requirements at this
position are significantly affected by the amino acid at
position 6 [25] These results suggested that the
combina-tion of amino acids at posicombina-tions 3 and 6 might be a critical
determinant for protein N-myristoylation In this study, to
examine the effect of the combination of amino acids at
positions 3 and 6 on protein N-myristoylation, sequential
vertical-scanning mutagenesis of the amino acids at
positions 3 and 6 in a model N-terminal sequence was
performed and the susceptibility of these mutants to
protein N-myristoylation was evaluated by metabolic
labeling in an in vitro translation system using rabbit
reticulocyte lysate
Experimental procedures
Materials Restriction endonucleases, DNA-modifying enzymes, RNase inhibitor, and Taq DNA polymerase were pur-chased from Takara Shuzo (Japan) The mCAP RNA capping kit and proteinase K were from Stratagene RNase was purchased from Boehringer-Mannheim (Germany) Rabbit reticulocyte lysate was from Promega [3H]leucine, [3H]myristic acid, [35S]methionine and Amplify were from Amersham (UK) The Dye Terminator Cycle Sequencing kit was from Applied Biosystems Anti-human TNF polyclonal Ig was purchased from R & D systems Pro-tein G Sepharose was from Pharmacia Biotech Other reagents purchased from Wako Pure Chemical, Daiichi Pure Chemicals, and Seikagaku Kogyo (Japan) were of analytical or DNA grade
Plasmid construction Plasmid pBluescript II SK(+) lacking ApaI and HinDIII sites was constructed as described previously [28], and designated pB Plasmid pBDpro-TNF, which contains a cDNA coding for the mature domain of TNF, was constructed as described [28,29] Plasmid pBMA(9)-TNF was constructed by utilizing PCR For this procedure, pBDpro-TNF served as a template, and two oligonucleo-tides [MA(9), B1] as primers (Table 1) After digestion with BamHI and PstI, the amplified product was subcloned into
pB at the BamHI and PstI sites Plasmids pBMGA(8) and pBMG6Swere constructed by a method similar to that used to construct pBMA(9)-TNF using two primers [MGA(8) and MG6S, respectively] as mutagenic primers (Table 1) The cDNAs coding for MG6X-TNF, in which Ser at position 6 in MG6S-TNF was replaced with each of the 19 other amino acids, were constructed by using a degenerated primer, MG6X, as mutagenic primer After digestion with BamHI and PstI, the amplified product was subcloned into pB at the BamHI and PstI sites The DNA
Table 1 Nucleotide sequences of oligonucleotides used for the construction of mutant TNF cDNAs N, A + C + G + T; K, T + G.
Trang 3sequences of the obtained plasmids were determined by the
dideoxynucleotide chain termination method and plasmids
having a distinct triplet codon corresponding to each of the
19 amino acids at position 6 were obtained The cDNAs
coding for MG3X-TNF, in which the Ala at position 3 in
MGA(8)-TNF was replaced with each of the 19 other
amino acids, were constructed by a method similar to that
used to construct MG6X-TNF using a degenerated primer,
MG3X, as a mutagenic primer pBGi1a-, pBGi1a-C3K- and
pBhippocalcin-TNF were constructed as described
previ-ously [25] The cDNAs coding for MG3X6S-,
MG3X6T-and MG3X6F-TNF, in which the Ala at position 6 in
MG3X-TNF was replaced with Ser, Thr and Phe,
respect-ively, were constructed as follows pB Gi1a-D1-12-TNF, in
which the DNA sequence encoding the N-terminal 12
amino acids of the mature domain of TNF was deleted
from pBGi1a-TNF, was first generated from pBGi1a-TNF
by PCR For this procedure, pBDpro-TNF served as a
template and two oligonucleotides (XHO-TNF13, B1) as
primers After digestion with XhoI and PstI, the amplified
product was subcloned into pBGi1a-TNF at the XhoI and
PstI sites DNA fragments coding for the N-terminal
10 residues of MG3X6S-, MG3X6T-, and MG3X6F-TNF
were amplified by PCR In this case, pBMG3X-TNF
served as a template and two oligonucleotides (T3 plus
6S-XHO, T3 plus 6T-6S-XHO, or T3 plus 6F-6S-XHO, respectively)
as primers After digestion with SacI and XhoI, the
amplified product was subcloned into pB Gi1
a-D1-12-TNF at the SacI and XhoI sites In these three sets of a-D1-12-TNF
mutants, the amino acids at positions 11 and 12 were
changed from Asp-Lys to Leu-Glu because of the insertion
of the Xho I-linker sequence pBGi1a-C3K-A7K-TNF, in
which the Ala at position 7 in pBGi1a-C3K-TNF was
replaced with Lys, was generated from pBGi1a-C3K-TNF
by PCR In this case, pBGi1a-C3K-TNF served as a
template and two oligonucleotides (C3K-A7K, B1) as
primers After digestion with BamHI and PstI, the
ampli-fied products were subcloned into pB at the BamHI and
PstI sites pBhippocalcin-K7A-TNF, pBMG3K6S-TNF,
and pBMG3K6S7K-TNF were constructed by a method
similar to that used to construct pBGi1a-C3K-A7K-TNF
The mutagenic primers used to construct these three
mutants were HC-K7A, MG3K6S, and MG3K6S7K,
respectively (Table 1) The DNA sequences of these
recombinant cDNAs were confirmed by the
dideoxynucleo-tide chain termination method [30]
In vitro transcription and translation
Methods essentially identical to those described previously
were employed [28] T3 polymerase was used to obtain
transcripts of these cDNAs subcloned into pB vector These
transcripts were purified by phenol/chloroform extraction
and ethanol precipitation prior to use Subsequently, the
translation reaction was carried out using the rabbit
reticulocyte lysate (Promega) in the presence of [3H]leucine,
[35S]methionine or [3H]myristic acid under conditions
recommended by the manufacturer The mixture
(com-posed of 17.5 lL of rabbit reticulocyte lysate, 0.5 lL of
1 mM leucine- or methionine-free amino acid mixture, or
1 mM complete amino acid mixture, 4.0 lL of [3H]leucine
(5 lCi), [35S]methionine (1 lCi), [3H]myristic acid (25 lCi)
or [3H]acetyl-CoA (2 lCi) and 3.0 lL of mRNA) was incubated at 30C for 90 min
Transfection of COS-1 cells and determination
of N-myristoylated proteins The simian virus 40-transformed African Green monkey kidney cell line, COS-1, was maintained in Dulbecco modified Eagle’s medium (DMEM, Gibco BRL) supple-mented with 10% fetal bovine serum (Gibco BRL) Cells (2· 105) were plated onto 35 mm-diameter dishes 1 day before transfection The pcDNA3 construct (2 lg; Invitro-gen) containing mutant TNF cDNA was used to transfect each plate of COS-1 cells along with 3 lL of LipofectAmine (2 mgÆmL)1; Gibco BRL) in 1 mL of serum-free medium After incubation for 5 h at 37C, the cells were re-fed with serum-containing medium and incubated again at 37C for
24 h The cells were then washed twice with 1 mL of serum-free DMEM and incubated for 5 h in 1 mL of DMEM with 2% fetal bovine serum containing [3H]myristic acid (100 lCiÆmL)1) Subsequently, the cells were washed three times with Dulbecco’s phosphate-buffered saline (DPBS) and collected with cell scrapers, and then lysed with 200 lL
of RIPA buffer [50 mMTris/HCl (pH 7.5), 150 mMNaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, proteinase inhibitors] on ice for 20 min The cell lysates were centrifuged at 21 000 g at 4C for 15 min in a microcentrifuge (HITACHI-CF15D2) and supernatants were collected After immunoprecipitation with anti-TNF
Ig, the samples were analyzed by SDS/PAGE and fluoro-graphy
Western blotting TNF samples immunoprecipitated from in vitro translation products or total cell lysates of each group of transfected cells were resolved by 12.5% SDS/PAGE and then transferred
to an Immobilon-P transfer membrane (Millipore) After blocking with nonfat milk, the membrane was probed with a specific goat anti-hTNF Ig as described previously [31] Immunoreactive proteins were specifically detected by incu-bation with horseradish peroxidase-conjugated anti-goat IgG (Santa Cruz) The membrane was developed with ECL Western blotting reagent (Amersham Corp.) and exposed to
an X-ray film (Kodak) Quantitative analysis of immuno-reactive proteins was carried out by scanning the X-ray film using an imaging densitometer (Bio-Rad GS-700)
Immunoprecipitation Samples containing TNF mutants were immunoprecipit-ated with a specific goat anti-hTNF polyclonal Ig (R & D systems) as described [28]
SDS/PAGE and fluorography Samples were denatured by boiling for 3 min in SDS/ sample buffer and then analyzed by SDS/PAGE on a 12.5% gel Thereafter, the gel was fixed and soaked in AmplifyTM (Amersham) for 30 min The gel was dried under vacuum and exposed to X-ray film (Kodak) for an appropriate period Quantitative analysis of the labeled
Trang 4proteins was carried out by scanning the fluorogram using
an imaging densitometer (Bio-Rad GS-700)
Results
Effect of the amino acid residue at position 6
in the N-myristoylation consensus motif on the efficiency
of the cotranslational N-myristoylation reaction
To determine the amino-terminal sequence requirements for
protein N-myristoylation, the relative roles of amino acids
in the N-myristoylation consensus motif, especially those
at positions 3 and 6, in protein N-myristoylation were
evaluated by metabolic labeling of model substrate proteins
in an in vitro translation system In this case, to avoid the
effects of amino acids at other positions in the N-terminal
region on protein N-myristoylation, MA(9)-TNF, a TNF
mutant in which 9 amino acids following the initiating Met
were changed to Ala, was used as the starting sequence to
evaluate the roles of the amino acids at distinct positions in
protein N-myristoylation (Fig 1)
As shown in Fig 2A (lane 2), translation of an mRNA coding for MA(9)-TNF in the presence of [3H]leucine gave rise to two translation products; one was the major product with an expected molecular mass (17 kDa) and the other was a fainter band with a molecular mass 2 kDa larger than expected No incorporation of [3H]myristic acid was detected in these translation products, as shown in Fig 2A (lane 10) [35S]Met labeling of this mutant revealed that [35S]Met was specifically incorporated into the upper of the two bands detected with [3H]Leu (lane 6) As there is no Met residue in the mature domain of TNF, this result indicates that the upper band corresponds to the protein species retaining the initiating Met residue and the lower band to the one lacking this residue When Ala at position 2 was changed to Gly, the obtained mutant [MGA(8)-TNF] was efficiently N-myristoylated, as shown in lanes 3 and 11
Fig 1 Schematic representation of generation of MA(9)-, MGA(8)-,
and MG6S-TNF cDNA coding for Dpro-TNF, which contains the
mature domain of TNF, was first generated from pro-TNF cDNA by
deleting the nucleotide sequence encoding the propeptide region of
pro-TNF Subsequently, cDNAs of MA(9)-, MGA(8)-, and
MG6S-TNF were generated from Dpro-MG6S-TNF cDNA by site-directed
muta-genesis.
Fig 2 MGA(8)-TNF with an N-terminal sequence MGAAAAAAAA
is N-myristoylated The mRNAs encoding G i1 a-, MA(9)-, MGA(8)-, and MG6S-TNF were translated in vitro in the presence of [ 3 H]leucine, [35S]methionine or [3H]myristic acid using rabbit reticulocyte lysate Following immunoprecipitation with anti-TNF Ig, the labeled trans-lation products were analyzed by SDS/PAGE and fluorography (A) The cDNAs encoding G i1 a-, MA(9)-, MGA(8)-, and MG6S-TNF were transfected into COS-1 cells, and their expression and N-myris-toylation were evaluated by Western blotting analysis and [ 3 H]myristic acid-labeling, respectively (B).
Trang 5The levels of incorporation of [3H]Leu and [3H]myristic
acid into the lower band of the expressed MGA(8)-TNF
were comparable with those into Gi1a-TNF [32], in which
the N-terminal 10 residues of the Gi1a protein were linked to
the N-terminus of the mature domain of TNF (lanes 1, 3,
9 and 11) These results revealed that MGA(8)-TNF is
efficiently N-myristoylated, similarly to a protein having a
natural N-myristoylation motif When Ala at position 6
in MGA(8)-TNF was changed to Ser, a similar level of
[3H]myristic acid incorporation was observed with the
obtained mutant (MG6S-TNF), as shown in lanes 4 and 12
The results obtained with MGA(8)- and MG6S-TNF
clearly indicate that Ser at position 6 is not critical for
protein N-myristoylation When these four mutants were
expressed in COS-1 cells and their susceptibility to protein
N-myristoylation was evaluated by in vivo metabolic
labeling with [3H]myristic acid, efficient protein
N-myris-toylation was detected equally with Gi1a-, MGA(8)- and
MG6S -TNF, as also observed in an in vitro translation
system (Fig 2B) In this case, the upper bands detected in
the in vitro translation system were not detected in the
Western blotting analysis of the expressed proteins in
COS-1 cells These results suggest that the incorporation of [3H]myristic acid into the major protein band expressed in the in vitro translation system fully reflect the in vivo protein N-myristoylation that occurs in intact cells
To determine the effect of the amino acid residue at position 6 in the N-myristoylation consensus motif on protein N-myristoylation, vertical scanning mutagenesis of the amino acid at position 6 in MG6S-TNF was performed;
a series of TNF mutants in which the Ser at position 6 in MG6S-TNF was changed to each of the 19 other amino acids was generated Subsequently, the susceptibility of these mutants to cotranslational protein N-myristoylation was evaluated by using the in vitro translation system The results for the 20 amino acids are arranged according to their radius of gyration All of these mutants, except for a mutant having a Cys residue at position 6, were efficiently expressed as determined by the incorporation of [3H]Leu, as shown in the upper panels of Fig 3A The labeling with [3H]myristic acid revealed that in addition to Ser and Thr, five other amino acids (Gly, Ala, Leu, Ile and Phe) were permitted at position 6 to direct efficient protein N-myristoylation (Fig 3B) In these mutants, a low level
Fig 3 Effect of the amino acid residue at position 6 in N-myristoylation consensus motif on the efficiency of cotranslational N-myristoylation reaction The mRNAs encoding MG6X-TNF were translated in vitro in the presence of [ 3 H]leucine or [ 3 H]myristic acid using rabbit reticulocyte lysate Following immunoprecipitation with anti-TNF Ig, the labeled translation products were analyzed by SDS/PAGE and fluorography Results for the
20 amino acids were arranged according to their radius of gyration Three independent experiments showed similar labeling patterns (A) The efficiency of protein N-myristoylation ([ 3 H]myristic acid incorporation/[ 3 H]leucine incorporation) of MG6X-TNF was compared by quantitative analysis of the fluorograms of [3H]myristic acid- and [3H]leucine-labeled proteins shown in the lower and upper panels of (A) Relative N-myristoylation efficiency of each MG6X-TNF was expressed as the percentage of the N-myristoylation efficiency of MG6L-TNF Results for the
20 amino acids were arranged according to their radius of gyration (B) ND, not determined.
Trang 6of [3H]myristic acid incorporation was detected with
mutants having Pro, Asn, Gln, Glu, His, Met, Tyr and
Trp at this position
It is generally accepted that Ser or Thr is preferred at
position 6 for protein N-myristoylation In fact, when the
number of each amino acid residues located at position 6 in
78 N-myristoylated proteins in which N-myristoylation was
experimentally verified listed in a recent report [33] were
counted, 74% (58 of 78) of these proteins had a Ser residue
and 13% (10 of 78) had a Thr residue at position 6 (Fig 4)
The number of N-myristoylated proteins having other
amino acids at position 6 accounted for only 13% (10 of 78)
of the total N-myristoylated proteins These observations,
taken together with the fact that five other amino acids
could be permitted at this position in the model substrate
protein, suggest that the presence of Ser or Thr at
posit-ion 6 might affect the amino acid requirements at other
positions, thereby favoring the susceptibility to protein N-myristoylation
Effect of the amino acid residue at position 6 in the N-myristoylation consensus motif on the amino acid requirement at position 3 for cotranslational protein N-myristoylation
In a previous report, we demonstrated that the amino acid
at position 3 strongly affected protein N-myristoylation, and the amino acid requirements at this position were significantly affected by the amino acid at position 6 [25] Therefore, we next determined the effect of the amino acid
at position 6 on the amino acid requirements at position 3 for protein N-myristoylation by vertical scanning muta-genesis We first determined the effect of the amino acid residue at position 3 in MGA(8)-TNF on protein N-myr-istoylation A series of TNF mutants (MG3X6A-TNF) in which Ala at position 3 in MGA(8)-TNF was changed to 19 other amino acids were generated and their susceptibility to cotranslational protein N-myristoylation was evaluated in the in vitro translation system The results revealed that the amino acid at position 3 in MGA(8)-TNF strongly affected protein N-myristoylation and only three amino acids (Ala, Asn and Gln) could direct efficient protein N-myristoyla-tion, as shown in Fig 5B In these mutants, a low level of [3H]myristic acid incorporation was detected with mutants having Ser, Cys, Val or Ile at this position Metabolic labeling of the same set of mutants with [3H]acetyl CoA revealed that efficient protein N-acetylation was detected in mutants having Ser, Thr, Asp, Glu or Met at position 3, as shown in Fig 5C These results indicate that the experi-mental results obtained by metabolic labeling with [3 H]my-ristic acid in the in vitro translation system do not reflect a simple enzyme reaction mediated by NMT, but do reflect the result of the overall reaction involving a set of cotranslational N-terminal modifications
When the Ala at position 6 in MG3X6 A-TNF was changed to Ser to generate MG3X6S-TNF, a dramatic change in the amino acid requirement at position 3 was
Fig 4 Amino acid residues at position 6 in naturally occurring
N-myristoylation motif The numbers of each amino acid residue
located at position 6 in 78 N-myristoylated proteins in which
N-myristoylation was experimentally verified listed in a recent report
[33] were counted and arranged according to their radius of gyration.
Fig 5 Effect of the amino acid residue at position 3 on the protein N-myristoylation and N-acetylation of MG3X6A-TNF The mRNAs encoding MG3X6A-TNF were translated in vitro in the presence of [3H]leucine (A), [3H]myristic acid (B) or [3H]acetyl CoA (C) using rabbit reticulocyte lysate Following immunoprecipitation with anti-TNF Ig, the labeled translation products were analyzed by SDS/PAGE and fluorography Results for the 20 amino acids were arranged according to their radius of gyration.
Trang 7observed: 11 amino acids (Gly, Ala, Ser, Cys, Thr, Val, Asn,
Leu, Ile, Gln, His) were permitted at position 3 to direct
efficient protein N-myristoylation, as shown in Fig 6A
A low level of [3H]myristic acid incorporation was also
detected with mutants having Pro, Asp, Glu and Met at this
position To determine whether the remarkable change in
the amino acid requirement at position 3 was specific for
Ser or not, the Ala at position 6 in MG3X6 A-TNF was
changed to other amino acids, and their susceptibility to
protein N-myristoylation was evaluated In this case, we
chose Thr and Phe to further analyze the effect of the amino
acid residue at position 6 on the amino acid requirement at
position 3 because of the presence of these amino acids at
position 6 in naturally observed N-myristoylated proteins
(Fig 4) The results revealed that amino acid requirements
at position 3 very similar to those of MG3X6A-TNF were
observed with both of these two series of mutants
(MG3X6T- and MG3X6F-TNF) as shown in Fig 6B,C
In these mutants, a low level of [3H]myristic acid
incorpor-ation was detected with several amino acids: Ser, Thr, Val,
Ile in MG3X6T-TNF and Thr, Val, Ile in MG3X6F-TNF
Thus, it was concluded from these observations that the
combination of amino acids at positions 3 and 6 constitutes
a major determinant for the susceptibility to protein
N-myristoylation When Ser was located at position 6, 11 amino acids (Gly, Ala, S er, Cys, Thr, Val, Asn, Leu, Ile, Gln, His) were permitted at position 3 to direct protein N-myristoylation When Ser was not located at position 6, only 3 amino acids (Ala, Asn, Gln) were permitted at position 3 to direct efficient protein N-myristoylation
The presence of a Lys residue at position 7
in the N-myristoylation consensus motif affects the amino acid requirement at position 3 and Lys becomes permitted at this position
We next determined whether the effect of the amino acid at position 6 on the amino acid requirements at position 3 found in the model substrate proteins were applicable to naturally N-myristoylated proteins or not The numbers of each amino acid residue located at position 3 in 74 naturally N-myristoylated proteins having Ala, Ser, Thr or Phe at position 6 listed in a recent report [33] were counted and are summarized in Fig 7 As shown in the figure, 95 per cent (70 out of 74) of these proteins had amino acid residues
at position 3 that were consistent with the amino acid requirements at position 3 for protein N-myristoylation found in this study All of the proteins in which the amino
Fig 6 Effect of the amino acid residue at position 6 in N-myristoylation consensus motif on the amino acid requirements at position 3 for cotrans-lational protein N-myristoylation The mRNAs encoding MG3X6S-, MG3X6T-, and MG3X6F-TNF were translated in vitro in the presence of [3H]leucine or [3H]myristic acid using rabbit reticulocyte lysate Following immunoprecipitation with anti-TNF Ig, the labeled translation products were analyzed by SDS/PAGE and fluorography Results for the 20 amino acids were arranged according to their radius of gyration A, B and C show results with MG3X6S-, MG3X6T-, and MG3X6F-TNF, respectively.
Trang 8acid at position 3 is inconsistent with our present results
have a Lys residue at this position These observations
suggest that an N-myristoylation motif having a Lys residue
at position 3 might have other specific structural
determi-nants that permit the Lys residue at position 3 while still
directing protein N-myristoylation When the N-terminal
sequences of five N-myristoylated proteins having a Lys
residue at position 3 listed in a recent review [4] were
compared, a striking similarity was observed; the amino
acid at position 7 was in all cases Lys (Table 2)
It was speculated from these observations that the specific
determinant that permits the Lys residue at position 3 might
be the Lys residue at position 7 To test this possibility, the effect of a Lys residue at position 7 on the amino acid requirement at position 3 was evaluated by using several TNF mutants
When the Cys residue at position 3 in Gi1a-TNF, which has a natural N-myristoylation motif at the N-terminus, was changed to Lys, N-myristoylation was significantly reduced,
as shown in Fig 8A lanes 1 and 2 However, when the Ala residue at position 7 of this mutant (Gi1a-C3K-TNF) was changed to Lys, efficient N-myristoylation was observed with the obtained mutant (Gi1a-C3K-A7K-TNF), as shown
in lane 3 In contrast, when the Lys residue at position 7 in hippocalcin-TNF, which has Lys residues at positions 3 and
7, was replaced with Ala, N-myristoylation was completely inhibited, as shown in lanes 4 and 5 These results clearly suggest that the specific determinant that permits the Lys residue at position 3 is the Lys residue at position 7
To further confirm this idea, the effect of the Lys residue
at position 7 on the amino acid requirement at position 3 was evaluated by using MG6S-TNF as a model substrate When the Ala residue at position 3 in MG6S-TNF was changed to Lys, N-myristoylation was completely inhibited,
as shown in Fig 8B lanes 1, 2, 4 and 5 However, when the Ala residue at position 7 of this mutant (MG3K6S-TNF) was changed to Lys, efficient N-myristoylation was observed with the obtained mutant (MG3K6S-7K-TNF),
as shown in lanes 3 and 6 These results strongly support the idea that the specific determinant that permits the Lys residue at position 3 is the Lys residue at position 7
Discussion
Protein N-myristoylation is a cotranslational protein modi-fication catalyzed by an enzyme, N-myristoyl transferase (NMT) NMT is a member of the GCN5 acetyltransferase (GNAT) superfamily All family members catalyze the transfer of an acyl group from CoA to a primary amino group NMT can be distinguished from other GNAT family members on the basis of the remarkable diversity of its protein substrates For example, it was reported recently that the Arabidopsis thaliana genome encodes 437 known
or putative NMT substrates, accounting for 1.7% of all proteins [17]
S cerevisiae Nmt1p is the best studied of the known NMTs The X-ray structure of a binary complex of Nmt1p with bound myristoyl-CoA has been determined [34] A structure of a ternary complex of Nmt1p with a bound
Fig 7 The combination of the amino acid residues at position 3 and 6 in
naturally occurring N-myristoylation motif The numbers of each amino
acid residue located at position 3 in 74 naturally N-myristoylated
proteins having Ala, Ser, Thr or Phe at position 6 listed in recent report
[33] were counted and arranged according to their radius of gyration.
A, B, C and D show results with N-myristoylated proteins having Ala,
Ser, Thr and Phe at position 6, respectively Filled bars, amino acid
residue consistent with the amino acid requirements at position 3
found in this study; striped bars, amino acid residue inconsistent with
the amino acid requirements at position 3 found in this study.
Table 2 N-terminal sequence of N-myristoylated proteins having Lys residue at position 3 Amino acids at positions 3, 6 and 7 are in bold type.
Ca2+binding/EF hand proteins
Visinin-like protein 3 (M)G K QN SK LRPEVLQDL
ADP-ribosylation factor
Trang 9nonhydrolyzable myristoyl-CoA analogue
[S-(2-oxo)penta-decyl-CoA] and an octapeptide substrate has also been
defined [34] The Nmt1p fold consists of a saddle-shaped
b-sheet flanked by a helices There is pseudo-2-fold
symmetry The N-terminal half forms the
myristoyl-CoA-binding site The C-terminal half forms the bulk of the
peptide-binding site Each half has a fold similar to the core
structure of GNAT superfamily members [8]
Proteins destined to become N-myristoylated begin with
the sequence Met-Gly The initiating Met is removed
cotranslationally by methionine aminopeptidase and then
myristic acid is linked to Gly-2 via an amide bond by NMT
However, not all proteins with an N-terminal glycine are
N-myristoylated and the ability to be recognized by NMT
depends on the downstream amino acid sequence In
addition, proteins with an N-terminal glycine may also be
subjected to another cotranslational modification,
N-acety-lation
The precise substrate specificity of S cerevisiae Nmt1p
has been characterized mainly by using purified enzyme and
synthetic peptides derived from the N-terminal sequences
of known N-myristoylated proteins [1,14,15] Some amino
acid preferences have been observed at distinct positions
downstream of the N-terminal glycine [1,13,16] In general,
Ser or Thr is preferred at position 6, and an N-terminal
consensus motif such as Met-Gly-X-X-X-Ser/Thr- [13] has been defined In addition to the preference for Ser/Thr residues at position 6, positively charged residues (Lys or Arg) are known to be preferred at position 7 [1,16] Amino acid preference was also observed at position 3: charged residues, aromatic residues and Pro are not allowed at this position [5] These amino acid preferences were confirmed
by recent studies on the NMT1p structure as determined by X-ray crystallography [34,35] In these studies, the structure
of a ternary complex of Nmt1p with a bound nonhydro-lyzable myristoyl-CoA analogue [S-(2-oxo)pentadecyl-CoA] and an Arf2p-derived octapeptide substrate, GLYASKLA, has been defined to 2.5 A˚ resolution The determined structure allows identification of specific residues within NMT that account for the amino acid preference at positions 3, 6 and 7 of the peptide substrate Ser6 (Ser5 in peptide GLYASKLA), which is greatly preferred in N-myristoylated proteins, is H-bonded to the side chain of His221 in NMT1p, as well as the backbone amides of Asp417 and Gly418 in NMT1p Lys7 (Lys6 in peptide GLYASKLA), also preferred in N-myristoylated proteins,
is H-bonded to the side chains of Asp417 and Gly418 in NMT As for Leu3 (Leu2 in peptide GLYASKLA), it was shown that contacts between the side chain of Leu3 and pantetheine of myristoyl-CoA complete formation of the
Fig 8 The presence of a Lys residue at position 7 affects the amino acid requirement at position 3 and allows Lys to occur at this position mRNAs encoding G i1 a-, G i1 a-C3K-, G i1 a-C3K-A7K-, Hippocalcin-, Hippocalcin-K7A-, MG6S-, MG3K6S-, MG3K6S-7K-TNF were translated in vitro in the presence of [ 3 H]leucine or [ 3 H]myristic acid using rabbit reticulocyte lysate Following immunoprecipitation with anti-TNF Ig, the labeled translation products were analyzed by SDS/PAGE and fluorography (A) Results with G i1 a-, G i1 a-C3K-, G i1 a-C3K-A7K-, Hippocalcin-, Hippocalcin-K7A-TNF (B) Results with MG6S-, MG3K6S-, MG3K6S-7K-TNF.
Trang 10peptide-binding site and at the same time generate a 90
bend in the peptide backbone, turning it away from
myristoyl-CoA and toward a peptide-binding groove Thus,
the amino acid at position 3 is important for positioning the
substrate peptide in the peptide-binding site The residues
described above in NMT1p that interact with GLYASKLA
are highly conserved in NMTs derived from other species
These results indicate that in addition to the Gly at position
2, the amino acids at positions 3, 6, and 7 in the substrate
protein play important roles in substrate recognition by
NMT These findings were further confirmed by an
Ala-scanning mutagenesis study designed to define the extent to
which residues at positions 2, 3, 5, and 6 of GLYASKLA
contribute to proper placement of the N-terminal Gly in
the active site [35] In these experiments, a panel of
GLYASKLA derivatives with single Ala substitutions at
these positions was produced and presteady-state kinetic
analysis was performed The results revealed that Ala
substitution for Leu2, Ser5, or Lys6 produced a 12–18-fold
reduction in the burst rate Based on these results, it was
postulated that differences in the efficiency of
N-myristoy-lation of various cellular proteins may arise in part because
of differences in the presentation of Gly2 dictated by
interactions among the residues at positions 3, 6, and 7 of the
substrate and elements in the enzyme’s peptide-binding site
Thus, it is well established that in addition to the Gly
at position 2, amino acids at positions 3, 6, and 7 play
important roles in substrate recognition by NMT However,
the relative role of these residues in substrate recognition,
the relationship between amino acids that reside in these
three distinct positions, or favorable amino acid
combina-tions in these posicombina-tions are not yet well characterized
In a previous report, we showed that metabolic labeling
in an in vitro translation system is an effective strategy to
characterize the N-terminal sequence requirements for
cotranslational protein N-myristoylation Using this assay
system, we demonstrated that the amino acid residue at
position 3 strongly affects protein N-myristoylation, and the
amino acid requirements at this position were significantly
affected by the amino acid at position 6 [25] These results
suggest that the combination of amino acids at positions
3 and 6 might be a critical determinant for protein
N-myristoylation
In the present study, to examine the effect of the
combination of amino acids at positions 3 and 6 on protein
N-myristoylation, sequential vertical-scanning mutagenesis
of the amino acids at positions 3 and 6 in a model substrate
protein having a sequence MGAAAAAAAA at its
N-terminus was performed and the susceptibility of these
mutants to protein N-myristoylation was evaluated by
metabolic labeling in an in vitro translation system using
rabbit reticulocyte lysate The results revealed that the
combination of amino acids at positions 3 and 6 strongly
affected the susceptibility of the protein to protein
N-myristoylation When Ser was located at position 6, 11
amino acids (Gly, Ala, S er, Cys, Thr, Val, Asn, Leu, Ile,
Gln, His) were permitted at position 3 to direct efficient
protein N-myristoylation In contrast, when Ala, Thr or Phe
was located at position 6, only 3 amino acids (Ala, Asn,
Gln) were permitted at position 3 to direct efficient
modification These results clearly indicate that the amino
acid residues permitted at position 3 are affected by the
amino acid residue reside at position 6 The fact that the increase in the number of permitted amino acid residues at position 3 in response to varying the amino acid at position
6 was specific for the Ser residue well explains the fact that Ser-6 is frequently observed in the naturally observed N-myristoylated proteins The mechanism by which the Ser6 of the substrate affects the amino acid residue permitted at position 3 is not clear It is possible to speculate that this phenomenon is mediated by the specific interaction between Ser6 of the substrate and elements in the enzyme’s peptide-binding site This specific interaction probably induces the changes in the structure of the peptide-binding site that cause the alterations in the permitted amino acid at position 3 Unfortunately, the only reported structural information about interactions between an NMT and its peptide substrates comes from the ternary structure with bound Arf2p-derived GLYASKLA, which has a Ser residue
at position 6 Therefore, if we could obtain the X-ray structure of NMT bound to a peptide which does not have a Ser residue at position 6, it might be possible to elucidate the mechanism by which the Ser at position 6 significantly affects the amino acid residue permitted at position 3
In addition to the combination of amino acids at positions 3 and 6, it was demonstrated that the combination
of amino acids at positions 3 and 7 also affects the susceptibility of the protein to protein N-myristoylation In this case, the presence of Lys at position 7 was found to affect the amino acid requirement at position 3, and allowed Lys to occur at this position This finding demonstrates again that the specific interaction between an amino acid at
a distinct position in the substrate protein and elements in the enzyme’s peptide-binding site will induce changes in the structure of the peptide-binding site that cause an alteration
in the permitted amino acid at an other position In the present study, we focused our attention on the effect of amino acid residue at positions 6 and 7 of substrate protein
on the amino acid requirement at position 3 for protein N-myristoylation, and did not study the effect of amino acids
at other positions Recent studies revealed that, for complete substrate protein, at least the N-terminal 17 residues of the substrate protein experience amino acid type variability restrictions for protein N-myristoylation [33] Therefore, it might be possible that the amino acids located beyond position 7 would also affect the amino acid requirement at position 3 Further studies are required to fully characterize the favorable combinations of amino acids at distinct positions in the substrate protein
It is very important to determine whether the restrictions
on the amino acid combinations at positions 3, 6 and 7 in the substrate protein for protein N-myristoylation found in this study are applicable to NMTs derived from different species As described previously, the residues in NMT1p that interact with the amino acids at positions 2, 5 and 6 in the Arf2p-derived peptide GLYASKLA are highly con-served in NMTs derived from other species Therefore, favorable combinations of amino acids at positions 3, 6 and
7 in substrate protein for protein N-myristoylation might be similar in many of the NMTs in different species In fact, the amino acid requirements found in this study were fully consistent with the N-terminal sequence of 78 N-myristo-ylated proteins derived from various species in which N-myristoylation was experimentally verified