(Received for publication, August 5, 1996)
From the Department of Molecular Biotechnology, University of Washington, Seattle, Washington 98195-7730
The polymerase-associated phosphoprotein (P protein) from Sendai virus, a murine Paramyxovirus, is reported in the literature to be a highly phosphorylated protein. In vitro studies have detected phosphorylation in different regions of the protein, while a single phosphopeptide (identified as the sole phosphorylation) site) was observed using in vivo techniques. In this work, two phosphorylation sites of the P protein from Sendai virus are localized by a direct approach using matrix-assisted laser desorption ionization/quadrupole ion trap mass spectrometry. A computer-aided approach is used to confirm peptide identification.
Sendai virus is the murine prototype of the Paramyxovirus, belonging to the order of Mononegalvirales. Related human pathogens include types of parainfluenza virus, mumps, measles, and respiratory syncytial virus as well as the more distantly related Filoviridae viruses Marburg and Ebola. Animal pathogens comprise Newcastle disease, cattle and bird parainfluenza viruses, canine distemper, and murine pneumonia. An understanding of the functioning of the Sendai virus prototype would provide insights into the function and replication of these ubiquitous human and animal viruses.
The genome of the Sendai virus consists of a single strand of RNA with negative polarity that codes for at least six structural and five nonstructural proteins (1-3). The RNA core is encapsidated by a helical nucleocapsid protein (NP). Virions enter cells directly through surface membranes, and viral replication and transcription (mediated by viral polymerase) begin immediately in the cytoplasm. Polymerase activity is carried out by the polymerase-associated phosphoprotein (P protein,1 Mr = 65,000) and the large protein, L protein, (Mr = 200,000) (4, 5). Almost all of the viral proteins are phosphorylated, however, the P protein appears to be more heavily phosphorylated on a mol/mol basis (6, 7). The P protein also seems to be modular in nature (8). N-terminal and C-terminal domains are conserved among Paramyxoviruses (52% and 69% homology, respectively), while a 100-residue region in the middle is variable with ~11% homology (9, 10). The C-terminal domain has been shown to stabilize the L protein (11), and the N-terminal region interacts with the NP protein and has been shown to be essential for RNA encapsidation as well as RNA synthesis (12).
A number of studies were undertaken to locate sites of post-translational modification in order to understand the role that phosphorylation plays in the function of the P protein. In vitro experiments produced conflicting results where phosphorylation sites were detected in the first N-terminal quarter of the protein (7) or in the second N-terminal quarter of the protein (13). More recent work (14) showed that cell-free phosphorylation using virion-associated protein kinase as a phosphorylating agent caused phosphorylation of both serine and threonine. In contrast, intracellular experiments in which the phosphorylation state of the P protein was analyzed during virus replication indicated that phosphorylation occurred only on serine residues. The number of detected phosphorylation sites also differed between the in vitro and the in vivo techniques. Enzymatic digestion of the virion-associated protein kinase-phosphorylated P protein using trypsin followed by two-dimensional thin layer electrophoresis (TLE) produced four major spots and nine minor spots (14). Similar experiments utilizing intracellular analysis produced one major spot with several minor spots, and it was concluded that the P protein in infected cells was primarily phosphorylated at one adjacent site or a set of adjacent sites (14). Site-directed mutagenesis was subsequently used to identify the primary location of phosphorylation on the P protein (15). In a parallel effort described here, direct analysis using quadrupole ion trap mass spectrometry was employed to identify P protein phosphorylation sites.
Mass Spectrometry
A quadrupole ion trap mass spectrometer (Finnigan MAT, San Jose,
CA) was placed in the vacuum manifold of a TSQ 70 triple quadrupole
mass spectrometer (Finnigan MAT) and was interfaced to an external
matrix-assisted laser desorption ionization (MALDI) source as described
previously (16). A 200-µm core-fused silica optical fiber was used to
ionize the sample by irradiation with a 337-nm beam from a nitrogen
laser (Laser Science, Inc., Newton, MA). The ion trap volume was filled
with helium to an uncorrected pressure of 5 × 104
torr, then argon was added using a separate needle valve to bring the
final uncorrected pressure to 5.5 × 10
4 torr. The
addition of ~10% (pressure/pressure) argon to the trapping volume
serves to improve the trapping efficiency (17) and the resolution of
the mass spectrum and slightly increases the amount of fragmentation
achieved upon injection into an ion trap. The ion beam was focused into
the ion trap using a three-element einzel lens. Typical lens voltages
were as follows: lens 1,
5; lens 2,
190; lens 3,
20; ion gate,
200 (open)/+ 200 (closed); trap float,
11, probe held at 5 V.
The ion trap scan employed to obtain a single-stage mass spectrum
consisted of the following steps (16). A 100-ms delay was used to
adjust the laser cycle to ~4.5 Hz. Ions were allowed into the ion
trap during a 5-ms ionization period when the gating tube lens was set
to "open." The transistor-transistor logic signal used to trigger
the ion gate operation was also employed to trigger the laser to fire
~500 µs after the ion gate opened. The rf level during the
ionization period was typically set to 1154 V0p. This enabled ions with m/z
values above 100 Da to be stably trapped. The amplitude of the rf
voltage was ramped to 7500 V0
p in order to
eject matrix ions with m/z values less than 650 Da. The
electron multiplier was then turned on. Data acquisition was accomplished by ramping the amplitude of the rf signal while applying a
supplementary signal to the end cap electrodes to resonantly eject ions
through the end cap electrode and into the conversion dynode/electron
multiplier detection system (18, 19). Details appear in the relevant
figure legends. The conversion dynode and electron multiplier were set
at
15 kV and 1.2 kV, respectively, for data acquisition. Supplemental
signals were also applied to the end cap electrodes to cause
amplification of ion trajectories by resonance excitation (20).
Subsequent ion fragmentation was induced by collisions with the helium
damping gas and afforded amino acid sequence information. Mass spectra
were displayed on a Compaq 386 PC then ported to a DECStation 2100 for
data acquisition and analysis. Base-line subtraction at a given
threshhold was performed to remove artifacts due to software
normalization.
A LaserMAT (Finnigan MAT) linear time-of-flight (TOF) instrument was also used to screen fractions for the presence of phosphopeptides and to determine the number of peptides in each sample. Ions generated by the MALDI process were extracted from the source and accelerated into a field-free flight tube employing a 20-kV potential. Detection was accomplished using a 15-kV conversion dynode coupled to an electron multiplier.
Sample Preparation
Recombinant 32P-labeled P protein from Sendai virus
was infected into CV1 cells and purified by immunoprecipitation (15), then enzymatically digested with trypsin or chymotrypsin. A portion of
the protein digests were separated by two-dimensional TLE. Spots on the
gel were excised and peptides were extracted (14). Aliquots of both the
protein digests and the extracted peptides (estimated to contain 1-3
nmol of material) were lyophilized and reconstituted in 0.1%
trifluoroacetic acid (Aldrich) to a concentration of 0.1-1
mM. This material served as a stock solution and was stored
at 4 °C.
Aliquots of the stock solutions representing
an estimated 250 pmol of material were loaded onto a 20-µl Peek
injection loop and concentrated by washing with 100% Solvent A (0.1%
trifluoroacetic acid) on an -Chrom column (300-Å pore diameter
packed with Reliasil C18, 2 mm × 10 cm, Upchurch
Scientific, Oak Harbor, WA) using an ABI 140B dual pump solvent
delivery system (Applied Biosystems, Foster City, CA) pumping at a flow
rate of 75 µl/min. Peek materials were utilized instead of stainless
steel to minimize sample losses due to the interaction of
phosphopeptides with iron. Peptides were separated by reverse phase
high performance liquid chromatography (HPLC) and eluted utilizing a
60-min gradient of 0-80% Solvent B (70:30:0.085 acetonitrile (HPLC
grade, EM Science, Gibbstown, NJ):water:trifluoroacetic acid, v/v/v).
Peptides were detected by monitoring UV absorbance at a wavelength of
220 nm. Fractions were collected into polypropylene microcentrifuge
tubes, lyophilized to dryness (Savant Instruments, Farmingdale, NY) and
stored at
4 °C. Samples were subsequently reconstituted in 10-15
µl of 0.1% trifluoroacetic acid prior to analysis.
Manual Edman degradation was performed by reconstituting lyophilized fractions in 10 µl of 5% phenylisothiocyanate in pyridine added to 10 µl of 50% aqueous pyridine, followed by heating at 37 °C for 30 min. The organic layer was extracted twice with 20 µl of 2:1 heptane:ethyl acetate, and the aqueous material was lyophilized. Cleavage of the N-terminal amino acid residue was accomplished by adding 10 µl of trifluoroacetic acid to the sample, heating at 37 °C for 15 min, and then lyophilizing. A final extraction was made using 30 µl of water combined with 50 µl of n-butyl chloride. The sample was lyophilized and then reconstituted in 0.1% trifluoroacetic acid and applied to the probe tip as described below.
MALDIThe matrix consisted of a saturated solution of
-cyano-4-hydroxycinnamic acid (Aldrich) in an equivolume mixture of
0.1% trifluoroacetic acid and acetonitrile. An 0.5-µl aliquot of
sample was co-deposited onto a gold-plated stainless steel probe tip with 1 µl of matrix and allowed to air dry. The probe was then inserted into the vacuum chamber utilizing a ball valve, and the laser
was employed to ionize the sample.
Data Analysis
A data base consisting of the transcribed genetic sequence of
the P protein was constructed. The PEPMTM (Finnigan MAT)
algorithm (21) was used to identify all possible sequences of amino
acid residues in the data base with molecular masses within ±5 Da of
the experimentally determined molecular mass for the peptide of
interest. The software also listed the predicted y-, b-, and a-type
ions (nomenclature described in Ref. 22) to aid in the analysis of the
fragmentation mass spectra. SEQUEST, a data base-searching algorithm
developed in our laboratory (23), was employed to confirm the
identification of the amino acid sequences. A list of possible
sequences in the P protein data base was generated, and theoretical
fragmentation mass spectra were constructed. The experimental
fragmentation mass spectra were then compared to theoretical
fragmentation mass spectra, scored, and ranked. Peaks in the mass
spectra labeled with a "P" refer to fragmentation products
generated from a phosphorylated peptide. The notation "" refers
to the loss of a guanidino group from an Arg-containing ion, while
"*" refers to the loss of water or ammonia. Peaks labeled with
bnym refer to internal cleavage products from
fragmentation at Pro, His, or Arg residues, e.g. given a
peptide sequence LPQGW, b2y1 corresponds to
PQGW, b2y2 corresponds to PQG, etc.
Peptide Mapping
Purified Sendai virus P protein was enzymatically digested using
trypsin. Resultant peptides were separated by HPLC, fractions were
collected, and samples for mass spectrometric analysis were prepared as
described above. The MALDI ion trap and the MALDI TOF were used to
screen the fractions for phosphopeptides and to provide a peptide map
of the P protein. The ion trap was employed to obtain molecular mass
information. The TOF was used to determine the number of peptides in
each fraction since the ion trap mass spectra were complicated by the
presence of fragmentation products. Some degree of fragmentation of
peptide ions upon injection into ion trap mass spectrometers is common.
The decomposition provided by metastable decay using the matrix
-cyano-4-hydroxycinnamic acid affords abundant sequence-specific
fragmentation products that can be diagnostic for the structure of
known biological molecules (24-28). The sequence-specific
fragmentation produced upon injection into the ion trap provided
verification of peptide assignment without the need to perform an
additional tandem mass spectrometry experiment. Fig. 1
shows the amino acid sequence for the P protein and the sequence
coverage obtained by tryptic mapping. The tryptic map covered ~61%
of the protein by mass with most of the coverage on the N-terminal half
of the protein where the phosphorylation sites were expected (7, 13).
Trypsin digestion produces 71 expected peptides. Thirty-four peptides
were mapped, representing 48% of the expected trypsin fragments. The
molecular masses for 25 of the peptides not identified were below the
low-mass cutoff determined by the matrix ejection pulse and were not
detected. The remaining 12 peptides accounted for 26% of the sequence
and were not unambiguously identified. The P protein was also digested with chymotrypsin to identify phosphopeptides not observed in the
trypsin digest.
Trypsin Digestion
MALDI/Time-of-flight Mass SpectrometryOne trypsin-generated
peptide from HPLC fraction 16 with m/z 2911 was detected.
The mass spectrum illustrated in Fig. 2 shows detection
of two major peaks with m/z values of 2911 and 2991. These
differ in mass by 80 Da, corresponding to the addition of HPO3 to the hydroxyl group on the side chain of Ser. The
mass difference suggests that the lighter peptide is the
non-phosphorylated version of the heavier peptide. The mass corresponds
to that of the tryptic peptide-(255-282) and confirmation of the
assignment using the ion trap mass spectrometer is discussed below.
Four serine residues are contained within the sequence; thus the site of phosphorylation cannot be assigned based on the molecular mass data.
MALDI/Ion Trap Mass Spectrometry
Analysis of
trypsin-generated HPLC fraction 16 afforded the mass spectrum shown in
Fig. 3. The dominant peak in the spectrum was at
m/z 2909, correlating with the signal observed using the TOF
mass spectrometer. A small peak at m/z 2989 was observed. The mass difference of 80 Da again indicates that the heavier peptide
was probably the phosphorylated analog of the lighter peptide. A strong
signal is also observed at m/z 2890. This signal corresponds
both to a loss of ~98 Da (neutral loss of phosphoric acid from the
phosphopeptide at m/z 2989) and a loss of water from the
unphosphorylated peptide. This loss of 98 has been shown to be a
signature for the presence of phosphorylated serine and threonine
(28-30).
A search of the P protein sequence using PEPMTM identified 37 possible peptides with m/z values within ±5 Da of 2909. Of these, five contained Lys or Arg at the C terminus, corresponding to peptides produced by trypsin digestion. Only one of the five sequences corresponded to an expected tryptic fragment, residues 255-282. Theoretical values of b- and y-type ions for the sequence were calculated and compared to the fragment ions observed in the mass spectrum, shown in Fig. 3b). The observed fragment ions correspond to the product ions expected for the tryptic fragment-(255-282) (YNSTGSPPGKPPSTQDEHINSGDTPAVR) confirming the sequence assignment made using TOF analysis. The expected fragmentation products for the unphosphorylated sequence are displayed in Fig. 3a and observed ions are underlined. Unphosphorylated sequence ions corresponding to the major peaks in the mass spectrum were identified as were minor peaks arising from fragmentation of the phosphorylated species. The presence of a series of y- and b-type ions served to confirm the sequence assignment. Signal suppression of b-type ions is due to the large number of proline residues present. A number of signals were present corresponding to internal cleavages of the peptide modulated by proline and histidine residues. These peptides provided particularly strong signals when the C-terminal amino acid was Asp, in agreement with recent observations (28). The signal from the phosphorylated molecular ion was weak and a series of low abundance peaks afforded by fragmentation of the phosphopeptide were present. The presence of the y23 ion as well as its phosphorylated analog indicate that Ser-260 is the most likely site of phosphorylation. The phosphorylated analogs of y26-28 as well as b10, b14, b16, b19, b21-23, b26, and b28 also serve to confirm the identification. Contributions to low abundance signals can also arise from the presence of other co-eluting peptides. Site-directed mutagenesis experiments (15) have shown that Ser-Pro is a consensus sequence for phosphorylation of the P protein. The phosphopeptide generated using the trypsin digest contains four serine residues at Ser-257, Ser-260, Ser-267, and Ser-275. Only Ser-260 is followed by a proline; thus, the identification of Ser-260 as the phosphorylated residue would agree with the proposed mechanism of a proline-mediated kinase.
The SEQUEST data base-searching program was subsequently used to confirm the sequence assignment. Results in Fig. 3c indicate that the expected sequence was chosen as the first-ranked choice. The final ranking is determined by the value for Cn, and the answer is assumed to be reliable if the Cn value for the first choice is much greater than the Cn value for the second choice (23). This is the case for the sequence assignment in Fig. 3c; thus, the manually determined answer was verified by computer-aided interpretation. Additional stages of mass spectrometry were done on the most abundant fragmentation products (MS/MS and MS3); however, no additional sequence information was obtained. Subdigestion of the fraction on the probe tip using Asp-N was attempted but the resulting peptides were not detected. The same phosphorylated peptide sequence was identified using a sample extracted from the two-dimensional TLE experiment (data not shown).
The SEQUEST program is capable of searching sequences for post-translational modifications, such as phosphorylation. For the phosphopeptides identified in this work, the best results were obtained by searching using the unphosphorylated mass on a sequence data base where phosphorylation was not included. We infer from this that the facile loss of phosphoric acid from the phosphopeptide tends to suppress further fragmentation; therefore, fragmentation proceeds primarily from the unphosphorylated species. Similar effects have been observed on another MALDI-ion trap mass spectrometer (27).
Chymotrypsin Digestion
MALDI/Time-of-flight Mass Spectrometry, Fraction 13In
previous work (15), digestion with chymotrypsin and subsequent
two-dimensional TLE of the peptides produced two spots. In the present
work, two candidate phosphopeptides were also observed when screening
the fractions collected from the chymotrypsin digest. Two major peaks
of m/z 1738 and m/z 1807 from HPLC fraction 13 were detected in the mass spectrum displayed in Fig. 4.
The 69-Da difference between the peaks does not correlate well with the expected 80-Da mass difference between ions resulting from the phosphorylated and unphosphorylated forms of the peptide. However, subsequent analysis of this fraction by ion trap mass spectrometry revealed the presence of a co-eluting peak at m/z 1728, corresponding to the non-phosphorylated version of the peptide of
m/z 1807. The broadened peak in Fig. 4 that was assigned to
m/z 1738 results from the combined presence of
m/z 1728 and m/z 1743, a co-eluting peptide
identified as the chymotryptic fragment HIITDRGGKTDNTDSL, residues
429-444. The low resolution linear TOF mass spectrometer was not able
to discriminate between the peaks at that molecular mass. The peptide
of m/z 1728 corresponds to the chymotryptic fragment from
residues 240-255. Ser-249 is the only serine residue contained in this
sequence; thus, the site of phosphorylation was straightforward to
assign.
MALDI/Ion Trap Mass Spectrometry, Fraction 13
Analysis of
chymotrypsin-generated HPLC fraction 13 produced the mass spectrum
shown in Fig. 5b. The dominant peak in the mass spectrum was at m/z 1709. An abundant signal at
m/z 1807 correlated with that observed on the time-of-flight
instrument. A small peak at m/z 1728 was also observed. The mass
difference of 79 Da indicates that the peptide at m/z 1728 is the non-phosphorylated version of the peptide at m/z
1807. The mass difference of 98 Da between the peaks at m/z
1807 and m/z 1709 suggests that the lighter peptide results
from the facile loss of phosphoric acid from the heavier peptide as
well as from the loss of water from m/z 1728. The presence
of a low abundance signal from a co-eluting peak at m/z 1743 corresponding to residues 429-444 was also observed.The PEPMTM software was used to search the P protein data base
for amino acid sequences with a mass of 1728 Da. Of 36 possible
peptides identified, seven contained a C-terminal residue resulting
from chymotrypsin cleavage on the carboxyl side of tyrosine,
tryptophan, leucine, or phenylalanine. Five of these peptides
corresponded to expected chymotrypsin-generated fragments. One step of
manual Edman degradation was performed as described above, and the
resulting mass spectrum is shown in Fig. 6. The
mass-to-charge ratios of the ions shifted by approximately 100 Da,
e.g. 950 Da 850 Da
, 1487 Da
1388 Da
, 1667 Da
1567 Da
, and 1709 Da
1610 Da, indicating valine or
threonine as likely candidates for the N-terminal residue. Only two
peptides out of the five possibilities contained Val or Thr at the N
terminus. Expected values for b- and y-type ions were calculated and
compared with the experimental spectrum in Fig. 5b. The
sequence of the chymotryptic peptide TPATVPGTRSPPLNRY (residues
240-255) fit the experimental data, strongly suggesting that the
phosphorylation site was at Ser-249. This peptide was also generated as
the top-ranked choice by SEQUEST, shown in Fig. 5c. The
expected fragmentation products for the peptide sequence are displayed
in Fig. 5a, and observed ions are underlined. The
presence of a series of b- and y-type ions served to confirm the
sequence identification. Some signal suppression following Pro or Gly
was observed. A loss of 43 Da corresponding to the loss of the
guanidino group from b-type ions containing arginine was observed for
most of the b ions. Similar losses have been observed for the
Arg-containing model peptide angiotensin using this instrument. The
presence of phosphorylated analogs of b10 and
y7, as well as y8, y10,
b12, b14, and b16 afforded positive
identification of Ser-249 as the site of phosphorylation. The Edman
degradation results (Fig. 6) confirm the identification also with the
presence of the phosphorylated analogs of b9 and y7, as well as b11, b14, and
y13. An internal cleavage product RSPPLN, as well as its
phosphorylated analog, were also observed. Further confirmation for the
sequence identification was provided by analysis using electrospray
ionization on a TSQ700 triple quadrupole mass spectrometer (data not
shown) and by site-directed mutagenesis. The identification of the site
of phosphorylation agrees with the proline-mediated kinase mechanism
discussed above (15). The same results were obtained by analysis of
peptides extracted from one of the spots on the two-dimensional TLE gel
(data not shown).
MALDI/Time-of-flight Mass Spectrometry, Fraction 25
The mass
spectrum shown in Fig. 7 results from analysis of HPLC
fraction 25 and shows the presence of two major peaks. The signals at
m/z 2732 and m/z 2811 differ by 79 Da; thus, the
lighter peptide appears to be the nonphosphorylated version of the
heavier peptide. The mass of this peptide did not correspond to an
expected chymotryptic fragment. Further analysis by ion trap mass
spectrometry (see below) was used to assign the peptide to residues
228-253, a peptide overlapping that identified from HPLC fraction
13.
MALDI/Ion Trap Mass Spectrometry, Fraction 25
Mass
spectrometric analysis of chymotrypsin-generated HPLC fraction 25 afforded the mass spectrum shown in Fig. 8b.
The dominant peak in the spectrum was at m/z 2710 and an
abundant signal was also observed at m/z 2728, correlating
with data obtained by TOF methods. The phosphorylated analog appeared
with low abundance at m/z 2809, which was an 80-Da mass
difference from the peak at m/z 2728 and a 98-Da mass
difference from the peak at m/z 2710. Again, a facile loss
of phosphoric acid was observed.
PEPMTM was used to search the data base for all peptides with m/z values of 2730 Da ± 5 Da. The search produced 35 possible peptides. Of these, four sequences theoretically resulted from cleavage with chymotrypsin. Signals at 2616 and 2599 Da were used to generate possibilities for the largest b- and y-type ions. The largest y-type ion could correspond to a loss of 112 Da or 129 Da, indicating Leu, Ile, Asn, Lys, Gln, and Glu as possibilities. The largest b-type ion could correspond to a loss of 95 Da or 112 Da, indicating Leu, Ile, Asn, and Pro as possibilities. Sequences for all peptides with these possible terminating ions were investigated. Mass assignment errors in the ion trap can arise from the use of a two-point calibration rather than a five-point calibration for the extended mass range, as well as shifting caused by space-charge effects; thus, the mass windows considered were rather large. An MS/MS experiment (data not shown) indicated that the peaks at m/z 2175 and 1294 were from the precursor ion at m/z 2730. Expected sequence ions for 20 possible sequences were compared with the experimental mass spectrum. The sequence with the best fit corresponded to residues 228-253 (KRRPTNSGSKPLTPATVPGTRSPPLN). This peptide overlaps the phosphopeptide with m/z 1728 from residues 240-255. The expected fragmentation products appear in Fig. 8a, and observed signals are underlined. The presence of a series of b- and y-type ions were sufficient to verify the sequence assignment. As expected, fragmentation was suppressed in the vicinity of proline residues, and a signal from a number of internal cleavage products was observed. There are three serine residues found in this peptide. Since only peak differences of 80 Da were observed using the TOF, we assume that the overlap peptide was phosphorylated at only one site. The observation of phosphorylated analogs of b25, y25, y22-20, y18, y16-15, y13-12, y10, and y8 serve to confirm the site of phosphorylation on Ser-249 rather than on Ser-234 or Ser-236. In addition, only Ser-249 is followed by a proline residue, agreeing with the proposed kinase mechanism (15). The same results were obtained by analysis of peptides extracted from one of the spots on the two-dimensional TLE gel (data not shown).
Three phosphopeptides were observed using MALDI/quadrupole ion trap mass spectrometry. Digestion with trypsin produced one phosphopeptide corresponding to residues 255-282, while digestion with chymotrypsin produced two phosphopeptides that overlapped in sequence. The smaller peptide corresponding to residues 240-255 contained one serine residue at Ser-249. The larger overlap peptide corresponded to residues 228-253 with Ser-249 again identified as the site of phosphorylation.
In parallel work using site-directed mutagenesis (15), deletion mutants were screened to locate a putative region of phosphorylation. The dominant phosphorylation region corresponded to residues 238-254 containing one serine at Ser-249. Deletion of Ser-249 or mutagenesis in that region of the peptide resulted in more extensive phosphorylation of the minor phosphopeptides, particularly a minor phosphopeptide located between residues 253 and 316. Experiments to investigate the effect of deleting residues 253-316 on the overall state of phosphorylation were not carried out. Concurring with these results, we believe Ser-249 is indeed a dominant phosphorylated residue in the P protein. The question remains as to why the phosphorylated peptide fragment containing Ser-260 was observed using mass spectrometry and not detected, except as a minor species, by classical genetics experiments. One possibility is that the fragment was ionized particularly well by MALDI and that the other minor phosphopeptides were not observed because of low stoichiometries. Another possibility is that the structure of the region of the protein surrounding Ser-260 precluded efficient incorporation of labeling agents, reducing its detectability by autoradiography. More work is required to determine the exact location of the phosphorylation site of the peptide fragment.
The functional significance of P protein phosphorylation is unclear. Phosphorylation has been proposed to aid in multimerization of the P protein and the maintenance of its structural integrity (15), enabling viral replication and transcription. With the determination of the phosphorylation sites on the P protein, classical genetics can be employed to elucidate the precise role of phosphorylation in Sendai virus functioning.