Access to this full-text is provided by Springer Nature.
Content available from Nature Communications
This content is subject to copyright. Terms and conditions apply.
ARTICLE
Analysis of human acetylation stoichiometry
defines mechanistic constraints on protein
regulation
Bogi Karbech Hansen1, Rajat Gupta1, Linda Baldus2,3, David Lyon4, Takeo Narita 1, Michael Lammers2,3,
Chunaram Choudhary 1& Brian T. Weinert1
Lysine acetylation is a reversible posttranslational modification that occurs at thousands of
sites on human proteins. However, the stoichiometry of acetylation remains poorly char-
acterized, and is important for understanding acetylation-dependent mechanisms of protein
regulation. Here we provide accurate, validated measurements of acetylation stoichiometry at
6829 sites on 2535 proteins in human cervical cancer (HeLa) cells. Most acetylation occurs
at very low stoichiometry (median 0.02%), whereas high stoichiometry acetylation (>1%)
occurs on nuclear proteins involved in gene transcription and on acetyltransferases. Analysis
of acetylation copy numbers show that histones harbor the majority of acetylated lysine
residues in human cells. Class I deacetylases target a greater proportion of high stoichiometry
acetylation compared to SIRT1 and HDAC6. The acetyltransferases CBP and p300 catalyze a
majority (65%) of high stoichiometry acetylation. This resource dataset provides valuable
information for evaluating the impact of individual acetylation sites on protein function and
for building accurate mechanistic models.
https://doi.org/10.1038/s41467-019-09024-0 OPEN
1Department of Proteomics, The Novo Nordisk Foundation Center for Protein Research, Faculty of Health and Medical Sciences, University of Copenhagen,
Blegdamsvej 3B, DK-2200 Copenhagen, Denmark. 2Institute of Biochemistry, Synthetic and Structural Biochemistry, University of Greifswald, Felix-
Hausdorff-Str. 4, Greifswald 17487, Germany. 3Institute for Genetics and Cologne Excellence Cluster on Cellular Stress Responses in Aging-Associated
Diseases, CECAD, University of Cologne, Joseph-Stelzmann-Str. 26, 50931 Cologne, Germany. 4Disease Systems Biology Program, The Novo
Nordisk Foundation Center for Protein Research, Faculty of Health and Medical Sciences, University of Copenhagen, Blegdamsvej 3B, DK-2200
Copenhagen, Denmark. Correspondence and requests for materials should be addressed to C.C. (email: chuna.choudhary@cpr.ku.dk)
or to B.T.W. (email: brian.weinert@gmail.com)
NATURE COMMUNICATIONS | (2019) 10:1055 |https://doi .org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications 1
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Lysine N-ε-acetylation is a reversible protein posttranslational
modification (PTM) that was first identified on histones1.In
the past decade, sensitive mass spectrometry (MS) techni-
ques enabled identification of thousands of acetylation sites on
diverse cellular proteins2–4. Acetylation can be enzymatically
catalyzed by lysine acetyltransferases, however, recent data indi-
cates that acetylation also arises from nonenzymatic reaction with
acetyl-CoA5,6. Nonenzymatic acetylation potentially targets any
solvent accessible lysine residue, suggesting that nonenzymatic
acetylation sites are likely to greatly outnumber acetyltransferase-
catalyzed sites. As a result, enzyme-catalyzed acetylation is easily
overlooked within a vast background of nonenzymatic acetyla-
tion, presenting a needle-in-a-haystack problem for identifying
these sites. Proteome-wide analyses of lysine acetylation should
focus on identifying parameters that will help prioritize the
functional relevance of individual sites and provide mechanistic
insights. These parameters include regulation by acetyl-
transferases and deacetylases, dynamic turnover rates, and the
stoichiometry of modification. Regardless of the origin of acet-
ylation, enzyme-catalyzed or nonenzymatic, understanding the
stoichiometry of modification is important for determining the
impact of acetylation on protein function and for building
accurate mechanistic models.
We developed a quantitative proteomics method to determine
acetylation stoichiometry at thousands of sites by measuring
differences in the abundance of native and chemically acetylated
peptides6,7. We subsequently refined our method by incorporat-
ing strict criteria for accurate quantification of acetylated pep-
tides8. However, the stoichiometry of acetylation in human cells
remains poorly characterized.
Here we determine acetylation stoichiometry at thousands of
sites in human cervical cancer (HeLa) cells. We validate our
results using known quantities of peptide standards, using
recombinant acetylated proteins, and by comparison with acety-
lated peptide intensity. This high-confidence dataset is used to
calculate acetylation copy numbers in cells, to explore the rela-
tionship between stoichiometry and regulation by acetyl-
transferases and deacetylases, and to reveal mechanistic
constraints on protein regulation by acetylation.
Results
Measuring acetylation stoichiometry. We measured acetylation
stoichiometry in HeLa cells using partial chemical acetylation and
serial dilution SILAC (SD-SILAC) to ensure quantification
accuracy8(Fig. 1a). Two independent biological replicates were
performed, each using a different degree of chemical acetylation
and inverting the SILAC labeling between experiments. The
degree of chemical acetylation was estimated based on the median
reduction of unmodified peptides generated by tryptic cleavage at
one or two lysine residues (Supplementary Figure 1a). Based on
the estimated degree of chemical acetylation, we performed a
serial dilution of the chemically acetylated peptides to give
median ~1%, ~0.1%, and ~0.01% chemical acetylation. Acetylated
peptides were enriched and the differences between native
acetylated and chemically acetylated peptides quantified by MS
(Supplementary Data 1a). To ensure accurate quantification, we
required that the abundance of native acetylated peptides was
quantified by comparison with at least two different concentra-
tions of chemically acetylated peptides, and that the measured
SILAC ratios agreed with the serial dilution series. SILAC ratios
that did not follow the dilution series (allowing up to two-fold
variability) were defined as being inaccurately quantified, even
though one of the measurements may be correct. Quantification
error was reduced when the concentration of chemically acety-
lated peptides was most similar to native acetylated peptides
(Fig. 1b). However, quantification error was substantially higher
than in our previous experiments in bacteria8, likely due to the
greater complexity of the human proteome. The high error rates
highlight the need to control for quantification accuracy, and
show that comparing native acetylated peptides to just 1% che-
mically acetylated peptides results in a majority of false quanti-
fication (Fig. 1b). The measured stoichiometry of acetylated
peptides was significantly and highly correlated between inde-
pendent experimental replicates (Fig. 1c). The precision of our
measurements was also highly reproducible; the median ratio of
stoichiometry between replicates was 0.95, and 90% of the mea-
surements varied by less than a factor of two between replicates
(Fig. 1d).
For high stoichiometry (>10%) acetylation, the difference
between native acetylated and chemically acetylated peptides
becomes too small to accurately measure by SILAC quantifica-
tion, which is typically limited to differences greater than 2-fold
in magnitude. At 5% partial chemical acetylation, a peptide with
90% stoichiometry will have a SILAC ratio of 1.006, and a peptide
with 50% stoichiometry will have a SILAC ratio of 1.05
(Supplementary Figure 1b). These differences are too small to
accurately resolve, and can result in inaccurate stoichiometry
measurements that are out of bounds (greater than 100% or
negative). Due to the inherent limitations in calculating
stoichiometry for these peptides, we set a cutoff of maximum
10% stoichiometry, and classified all sites exceeding this cutoff as
having >10% stoichiometry. Only 16 peptides met these criteria,
and they harbored previously known high stoichiometry acetyla-
tion sites (Supplementary Data 1a), including; histones H3 K23
(2 different peptides), H3 K14, and H4 K16. Other peptides
harbored sites on acetyltransferases, such as CBP K1583
(2 different peptides) and N-alpha-acetyltransferase 50 K33, or
on proteins that catalyze reactions using acyl-CoAs, such as
hydroxymethylglutaryl-CoA lyase K48 and dihydroxyacetone
phosphate acyltransferase K643. These data show that, although
we are unable to accurately measure high stoichiometry
acetylation, we were able to identify these peptides, they represent
known or probable high stoichiometry acetylation sites, and they
constitute a small portion (0.2%) of the peptides analyzed.
Validating stoichiometry measurements. We used known
quantities of unmodified and acetylated peptide standards
(AQUA peptides) to determine stoichiometry directly at ten sites
on three proteins, cortactin (CTTN), nucleolin (NCL), and N-
acetyltransferase 10 (NAT10) (Supplementary Data 1b). Stoi-
chiometry determined by partial chemical acetylation (PCA) was
significantly correlated (r=0.94) with stoichiometry determined
using AQUA peptides (Fig. 1e). Furthermore, stoichiometry
measurements differed by a factor of two or less for a majority
(7/10) of the analyzed peptides (Fig. 1f). Three sites showed 3.4-,
6-, and 7-fold higher stoichiometry by PCA, indicating over-
estimation of stoichiometry by PCA or underestimation by
AQUA. These differences occurred at the site-level and were
therefore not attributable to errors in protein quantification.
We think that the agreement between these two methods is
notable when considering all possible sources of variability in
each measurement.
We further validated our measurements using two recombi-
nant, site-specifically acetylated proteins; malate dehydrogenase
(MDH2) K239ac and hydroxymethylglutaryl-CoA lyase
(HMGCL) K48ac. Recombinant acetylated proteins (SILAC-
light-labeled) were used as a spike-in standard to measure
acetylation stoichiometry in SILAC-heavy-labeled HeLa lysate.
Stoichiometry measured using two different concentrations of
recombinant acetylated protein (spike-in) agreed with our
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0
2NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
measurements using PCA, further supporting the accuracy of our
stoichiometry dataset (Fig. 1g).
Stoichiometry measurements were additionally validated by
comparison to acetylated peptide intensity that was corrected for
differences in protein abundance. We previously showed that
abundance-corrected intensity (ACI) is correlated to acetylation
stoichiometry in yeast6. ACI was significantly correlated with
acetylation stoichiometry in HeLa cells (Supplementary
a
b
eg
c
1%
10%
0.1%
0.01%
0.001%
1%
10%
0.1%
0.01%
0.001%
Log10 stoichiometry Exp.2
Log10 stoichiometry
Exp.1
1%
10%
0.1%
0.01%
0.001%
1%
10%
0.1%
0.01%
0.001%
Log10 stoichiometry
AQUA
Log10 stoichiometry
PCA
n = 3839, r = 0.89, P < 2e–16
n = 10, r = 0.94, P = 7e–5
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
1%
0.1%
0.01%
1%
0.1%
0.01%
Exp.1 Exp.2
Median % chemically acetylated peptides
Quantification error
Count
−3 −2 −1 0
Log2 stoichiometry
Exp.1/Exp.2
250
200
150
100
50
0
SILAC light (L )
Native acetylation (-nAc)
-nAc
SILAC heavy (H )
Chemical acetylation (-cAc)
-cAc
-cAc
cAc-
cAc-
cAc-
-cAc
-cAc
-cAc
Combine and
digest to peptides
Analyze data and
calculate stoichiometry
-nAc
-cAc
cAc-
-cAc
-cAc
-nAc
-cAc
cAc-
-cAc
-cAc
Antibody enrich
-Ac peptides
Quantify by MS
and serial dilution
1:1 dilution
m/z
1:10 dilution
m/z
1:100 dilution
m/z
SILAC
ratio H/L
= 7.0
SILAC
ratio H/L
= 0.7
SILAC
ratio H/L
= 0.07
2. Filter for accurate quantitation
3. Calculate stoichiometry
SILAC ratio ~ dilution series
MaxQuant
1. Computational analysis
S = stoichiometry
R = SILAC ratio cAc/nAc
C = % chemical Ac
S = C / (R - (1-C))
Experiment 1: SILAC L = Native acetylation; SILAC-H = Chemical acetylation; % chemical acetylation (C) = 3.5%
Experiment 2: SILAC L = Chemical acetylation; SILAC-H = Native acetylation; % chemical acetylation (C) = 10.4%
3
2
1
0
−1
−2
−3
Log2 ratio stoichiometry
PCA/AQUA
MDH2 K239ac
HMGCL K48ac
0.01 %
0.1%
Native
Native
Intensity
m/z m/z
Spike-in stoichiometry
0.015%
0.018%
PCA stoichiometry
1%
10%
Native
m/z m/z
Native
Intensity
Spike-in stoichiometry
7.8%
>10%
PCA stoichiometry
90%
321
f
d
Fig. 1 Measuring acetylation stoichiometry. aDiagram of the method used to measure acetylation (Ac) stoichiometry. bThe degree of quantification error
as determined by the fraction of SILAC ratios at each concentration of chemically acetylated peptides that was not consistent with SILAC ratios measured
in at least one different concentration of chemically acetylated peptides. cThe correlation between stoichiometry measured in independent experimental
replicates. The number of peptides (n), Pearson’s correlation (r), and P-value (P) of correlation are shown. dLow absolute variability between experimental
replicates. The histogram shows the distribution of Log2 ratios of stoichiometry in Experiment 1/Experiment 2 (Exp.1/Exp.2). eThe correlation between
stoichiometry measured using partial chemical acetylation (PCA) and absolute quantification (AQUA) peptide standards. fLow absolute variability
between stoichiometry measurements made by PCA and AQUA. gValidation of stoichiometry measurements using recombinant acetylated (100%)
proteins as a spike-in standard. Stoichiometry was measured at two different concentrations of spike-in protein (SILAC light, red) compared to SILAC
heavy-labeled HeLa (blue) for each acetylation site. Source data are provided as a Source Data file
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0 ARTICLE
NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Figure 1c), however, the predictive power of this correlation was
modest (r=0.48–0.52). There are several reasons for this modest
correlation. Firstly, peptide intensity is inherently variable.
Secondly, protein abundance estimates may be inaccurate. We
found that outlier data points using iBAQ-based protein
abundance were not outliers when using copy-number-based
protein abundance (Supplementary Figure 1c), indicating that
variability in protein abundance measurements contributes to
disagreement between ACI and stoichiometry measurements.
Thirdly, antibody-based acetylated peptide enrichment may be
peptide-sequence biased, which will introduce further variability.
Regardless of these limitations, ACI provides an easy method to
estimate the relative stoichiometry of acetylation sites, and the
significant correlation with our stoichiometry measurements by
PCA provides further support for the accuracy of our
measurements.
Copy number limits the detection of acetylated peptides. The
detection of acetylated peptides is biased to abundant proteins
(Fig. 2a). Furthermore, the fraction of lysines that are detected as
acetylated on any given protein is significantly correlated with
protein abundance (Fig. 2b). This bias is found in every acetylome
dataset that we have examined (Supplementary Figure 2a)9–12,
and indicates that acetylation occurs on most lysine residues in
cells and that protein abundance is a limiting factor in the
detection of acetylated peptides. These data further support the
notion that all solvent accessible lysine residues are acetylated to
some degree, either enzymatically or nonenzymatically.
Deep proteome measurements detect unmodified peptides
from proteins whose abundance spans seven orders of magnitude
(Fig. 2a)13. This raises the question of why we detect so few
acetylated peptides without antibody enrichment (Fig. 2c). The
signal intensity of acetylated peptides is comparable to unmodi-
fied peptides in the mass spectrometer (Supplementary Figure 2b,
c), indicating that we should be able to detect acetylated peptides
as readily as unmodified peptides. We compared copy numbers
for unmodified peptides to copy numbers for acetylated peptides
as determined from our stoichiometry measurements. The
distribution of acetylated peptide copy number shows that, in
the absence of acetylated peptide enrichment, most acetylated
peptides are at or below the detection limit of the mass
spectrometer, even in deep proteome measurements (Fig. 2d).
In contrast, acetylated peptides that were detected without
antibody enrichment occurred at copy numbers that were within
the detectable range of unmodified peptides (Fig. 2e). Thus, our
stoichiometry measurements are consistent with the inability to
detect acetylated peptides without enrichment. Strikingly, our
data indicate that some acetylation events are so rare that they
occur at a copy number that is less than one per cell (Fig. 2d).
Properties of high stoichiometry acetylation. We measured the
stoichiometry of acetylated peptides; however, individual acet-
ylation sites may occur on multiple different peptides due to
incomplete tryptic digestion, protein N-terminal acetylation, or
oxidized methionine residues. To examine acetylation stoichio-
metry at the site-level, we calculated the summed stoichiometry of
peptides containing the same acetylation site (Supplementary
Data 1c). This resulted in stoichiometry measurements for
6829 sites, with a median stoichiometry of just 0.02% (1/4000
molecules) (Fig. 3a). This represents very low levels of acetylation
for most sites, only 1% (66 sites) displayed stoichiometry >1%,
and ~15% (1014 sites) displayed stoichiometry >0.1%.
We performed UniProt keyword enrichment analysis to
examine the functional categories of proteins that are associated
with higher stoichiometry (>0.23% or >1%) acetylation (Fig. 3b).
Higher stoichiometry acetylation was overrepresented on nuclear
proteins involved in chromatin regulation and transcription. This
observation is consistent with the known nuclear functions of
acetyltransferases, deacetylases, and acetylated lysine-binding
bromodomain proteins. In fact, the keywords Bromodomain
and Acetyltransferase were significantly enriched in the group of
proteins with high stoichiometry acetylation. We were unable to
calculate stoichiometry for doubly acetylated peptides because of
the low frequency of chemical acetylation at both positions.
08
0
4
0
300
200
100
Log10 copy number Log10 copy number Log10 copy number
Log10 copy number
Count (proteins)
Count
(peptides)
Proteins (8425)
Ac proteins (2488)
ad
e
c
4325
100%
10%
1%
Log10 % lysines acetylated
on each protein
2488 Ac proteins
6753 Ac sites
r = 0.65, P < 2e−16
This study
Bekker-Jensen et al.
Proteins
11,264
10,118
Peptides
398,893
170,251
Ac peptides
732
282
% Ac peptides
0.18%
0.17%
Detection of acetylated peptides without antibody enrichment
−2 0
0
6000
5000
4000
3000
2000
1000
Count (peptides)
Peptides (158,796)
Ac peptides (7479)
246 2468
−202468
b
67
Fig. 2 Stoichiometry limits the detection of acetylated peptides. aAcetylation is biased to detection on abundant proteins. Protein copy number estimates
are from42.bThe fraction of acetylated lysines detected on any given protein is correlated with protein abundance. The scatterplot shows the %
lysines acetylated and copy numbers of 2488 acetylated proteins containing 6753 acetylation sites. The Pearson’s correlation (r), and P-value (P)of
correlation are shown. cThe number of peptides and acetylated peptides (Ac peptides) detected in deep proteome measurements from this study and13.
dThe distribution of peptide copy numbers from a deep proteome measurement and acetylated peptide copy numbers calculated from the peptide
stoichiometry and protein copy number. eThe distribution of acetylated peptide copy numbers for acetylated peptides that were detected without prior
antibody enrichment. Source data are provided as a Source Data file
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0
4NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
However, given the low stoichiometry of acetylation, doubly
acetylated peptides are unlikely to occur by random chance and
most likely reflect the activity of acetyltransferases. Consistent
with this idea, doubly acetylated peptides occurred on proteins
that were overrepresented for the same UniProt keywords that
were associated with high stoichiometry acetylation (Fig. 3b).
Thus, sites occurring on doubly acetylated peptides may occur at
high stoichiometry and are likely to be enzyme-catalyzed.
To investigate the relationship between stoichiometry and
subcellular compartmentalization we used immunofluorescence-
based protein localization as determined by the Human Protein
Atlas14. Acetylation stoichiometry was broadly distributed and
mostly similar in every subcellular compartment analyzed
(Fig. 3c). Mitochondrial acetylation occurred at a slightly, yet
significantly (P<5e
−5, Wilcoxon test), higher median stoichio-
metry. However, mitochondria contained the smallest fraction of
high (>1%) stoichiometry acetylation sites. In contrast, the
nucleus contained the greatest fraction of high stoichiometry
sites, which was approximately an order of magnitude greater
than in mitochondria (Fig. 3c).
We used IceLogo15 to determine whether high stoichiometry
acetylation was associated with neighboring amino acids.
Cysteine residues were notably overrepresented for sites with
>0.23% stoichiometry (10-fold higher than median stoichiome-
try), particularly in the −4, −3, and −2 positions (Fig. 3d).
However, this bias was absent when examining sites with >1%
stoichiometry, indicating that this overrepresentation was asso-
ciated with sites with moderately elevated stoichiometry.
Remarkably, sites with cysteine residues in the −4, −3, or −2
position constituted 35% (159/460) of the sites with >0.23%
stoichiometry. UniProt keyword enrichment analysis of the
proteins harboring these sites found a variety of enriched
keywords (Supplementary Data 1d). However, unlike high
stoichiometry acetylation in general (Fig. 3b), keywords describ-
ing processes associated with nuclear acetyltransferases, such as
Nucleus, Transcription, and Chromosome were notably absent.
These data suggest that cysteine residues may promote none-
nzymatic acetylation of downstream lysine residues, and these
sites constitute a substantial portion (35%) of sites with an
elevated (>0.23%) stoichiometry of acetylation. This conclusion is
a
0
400
300
200
100
Count (sites)
10%
100%
1%
0.1%
0.01%
0.001%
Stoichiometry
>0.23% stoichiometry
>0.23% stoichiometry
>1% stoichiometry
>1% stoichiometry
Doubly acetylated peptides
% associated with keyword
>2-fold keword enrichment
c
Cytosol
Nucleus or
nucleoplasm
Mitochondria
Vesicles
Plasma
membrane
n – 2745 2433 758 463 741
Median – 0.022% 0.025% 0.032% 0.023% 0.020%
>1% – 0.77% 1.40% 0.13% 1.08% 0.54%
0.001%
0.01%
0.1%
1%
10%
100%
Stoichiometry
25
–25
12
–12
Fold change
(P value < 0.05)
25
–25
12
–12
Fold change
(P value < 0.05)
d
–15 –13–12 –11 –9 –8 –7 –6 –5 –4 –3 –2 –1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15–10–14
–15 –13–12 –11 –9 –8 –7 –6 –5 –4 –3 –2 –1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15–10–14
b
Transcription regulation
Nucleosome core
Nucleotide−binding
Transferase
Acyltransferase
Isopeptide bond
Transcription
Methylation
Chromosome
RNA−binding
Nucleus
Ubl conjugation
Helicase
Acetylation
Chromatin regulator
DNA−binding Thiamine pyrophosphate
ATP−binding
Lyase
Bromodomain
Chromosome
Nucleus
Nucleosome core
Peroxisome
Nucleotide−binding
Acetylation
Chromatin regulator
Methylation
Helicase
Isopeptide bond
Bromodomain
Transcription regulation
Ubl conjugation
Citrullination
Activator
DNA−binding
Transcription
Coiled coil
Acyltransferase
Nucleus
Activator
Citrullination
Nucleosome core
Chaperone
DNA−binding
Methylation
Ribonucleoprotein
Glycolysis
Acetylation
Ubl conjugation
Isopeptide bond
Chromosome
Chromosomal rearrangement
Acyltransferase
Bromodomain
Coiled coil
50% 100%
10% 75%
25%
Fig. 3 Properties of high stoichiometry acetylation. aThe distribution of acetylation site stoichiometry for the 6829 sites measured in this study. bUniProt
keyword enrichment for the indicated classes of high stoichiometry acetylation sites (>0.23% and >0.1%) and for doubly acetylated peptides. The sizeof
the text is related to the fraction of sites associated with the keyword, and keywords that were more than two-fold enriched are colored red. cSubcellular
compartment analysis based on the Human Protein Atlas14. Category scatterplots show the distributions of acetylation site stoichiometry in each
subcellular compartment. The number (n) of sites analyzed, median stoichiometry (median), and percentage of sites with >1% stoichiometry (>1%) is
shown. dAmino acid sequence logos of the indicated classes of acetylation sites using IceLogo15. Source data are provided as a Source Data file
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0 ARTICLE
NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved
supported by a recent study that uncovered a similar bias for
higher stoichiometry acetylation at sites with proximal cysteine
residues16.
Histone acetylation. We measured stoichiometry at 57 histone
sites and found that high stoichiometry acetylation was mostly
restricted to sites on the N-terminal tails of core histones H2B,
H3, and H4 (sites on H2A were not measured) (Fig. 4a). The
stoichiometry of histone H3 and H4 acetylation sites has been
extensively studied17–20, and our measurements are comparable
to these previous measurements (Fig. 4b). However, our method
does not measure the stoichiometry of doubly acetylated peptides
that can arise from the lysine-rich N-terminal tails of the core
histones H2A, H2B, H3, and H4. Feller et al.18 found that the
stoichiometry of mono-acetylated histones H3 and H4 were more
abundant than di-acetylated H3 and H4, with the exception of H4
K5 +K12 and H3 K18 +23 (which is more abundant than K18
alone, but less abundant than K23 alone). We detected doubly
acetylated peptides containing the following sites on H2A (K5 +
K9, K9 +K11, K11 +K13), H2B (K5 +K11, K11 +K12, K15 +
K16, K16 +K20, K20 +K23, K34 +K43, and K116 +K120), H3
(K9 +K14, K18 +K23, and K27 +K36), and H4 (K8 +K12 and
K12 +K16). The stoichiometry of di-acetylated peptides remains
unexplored for H2B, and these peptides may be more abundant
than their mono-acetylated counterparts used to calculate stoi-
chiometry in this study. In addition, some acetylated peptides
from histone tails may not be detected because of their small size.
Thus, our estimates of histone acetylation stoichiometry may
underestimate the actual native stoichiometry at these positions
because these sites also occur on di-acetylated peptides or pep-
tides that we are unable to detect with our methodology.
Regardless, our data suggest that the stoichiometry of H2B
acetylation ranges from 0.5% to 5.6%. N-terminal H2B sites are
primarily acetylated by the CBP/p300 acetyltransferases12, which
also target H3K27 and H3K36. N-terminal H2B sites (K5, K11,
Non-histone
Histone (this study)
Histone (other)
H3K9
H3K14
H3K18
H3K23
H3K27
H3K36
H4K5
H4K8
H4K12
H4K16
0%6.
23.0%
3.0%
3.0%
6.0%
30.0%
0.9%
21.0%
10.7%
55.5%
0.3%
0.3%
9.8%
6.0%
14.9%
5.5%
5.3%
17.6%
8.9%
7.1%
13.3%
11.0%
27.8%
4.1%
25.1%
2.7%
6.9%
12.7%
35.7%
Histone
Ac site
Abshiru et al
>15%
>20%
>10%
This study
Feller et al
Zhou et al
Zheng et al
Ac copy number per cell
5e7
4e7
3e7
2e7
1e7
0
N-terminal
Histone Ac site Stoichiometry
H3.1 23 >20%
H3.1 14 >15%
H4 16 >10%
H2B 2-F 20
H2B 1-K 20
H1.4 17
5
5
5
5
5
5
2.1%
H2B 1-K 1.6%
H2B 1-M 20
H2B 1-M 0.98%
H2B 1-B 20 0.66%
H1.2 17 0.59%
H2B 1-D 0.43%
H2B 1-H 0.29%
H2B 1-L 0.26%
H2B 3-B 0.20%
H1.2 21 0.17%
H1.1 17 0.14%
C-terminal
Histone Ac site Stoichiometry
H2A.1 291 0.099%
H1.5 35 0.067%
H1.5 17 0.064%
H1.0 12 0.062%
H2A.1 303 0.061%
0.059%H1.2
H1.4 46 0.049%
H1.5 66 0.048%
0.039%H1.4 106
H1.4 63 0.036%
H1.5 49 0.034%
H1x 143 0.031%
H1.1 182 0.030%
H1.4 90 0.029%
H4 91 0.029%
H1.0 27 0.024%
H4 77 0.023%
H2A.1 115 0.022%
H2A.Z 115 0.021%
H2A 1-C 95
C-terminal
Histone Ac site Stoichiometry
H1.4 117
H2B 1-O 108 0.017%
0.017%
0.019%
H2B 1-O 120
H1.0 59 0.016%
0.015%
0.012%
0.012%
0.011%
0.011%
0.011%
0.010%
0.010%
0.009%
0.009%
0.009%
0.009%
0.009%
H2B 1-O 46
H2A 1 95
H1x 115
H2B 1-D 120
H2B 1-O 116
H1.5 78
H2B F-S 85
H2B 1-D 43
H2A 2-A 95
H1.0 52
H2B 1-K 120
H1.4 75
H1x 90
H1.0 69 0.008%
H1.0 40 0.007%
H1.0 82 0.006%
188
98
62
49
38
28
17
14
6
Short
Long
Exposure
H3 (15 kD)
H2A/H2B (14 kD)
H4 (11 kD)
26%
74%
16%
46%
38%
6e7
5.6%
2.9%
1.2%
168
0.019%
a
bcd
Fig. 4 Stoichiometry of histone acetylation. aThe diagram shows all histone acetylation sites whose stoichiometry was determined in this study. The sites
are ordered by descending stoichiometry. Note that high stoichiometry sites occur on the N-termini of core histones. bThe stoichiometry of histone
acetylation sites as determined in four independent studies17–20.cAn anti-acetylated lysine immunoblot of HeLa whole cell lysate. Cells were boiled in 2%
LDS to ensure histone extraction. Histones are annotated based on their expected molecular weight. dHistone acetylation sites constitute a majority of
acetylated lysine residues in cells. Stoichiometry and protein copy numbers were used to calculate the number of acetylated lysine residues for the
indicated classes of proteins. Source data are provided as a Source Data file
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0
6NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
K12, K15, and K16) show faster deacetylation kinetics compared
to K2012. Interestingly, the stoichiometry of K20 is greater than
K5, indicating that the lower stoichiometry at K5 is possibly due
to its more rapid turnover.
Histones are some of the most abundant and most highly
acetylated proteins in cells. Anti-acetylated lysine antibodies
prominently detect histone acetylation in western blots of whole
cell lysates, suggesting that histones harbor most of the acetylated
lysine residues in cells (Fig. 4c). Using our acetylated lysine copy
number estimates we found that histone acetylation accounted for
74% of the acetylated lysine residues in cells (Fig. 4d). If we
include histone sites whose stoichiometry was measured by
independent studies17–20, the fraction of acetylated lysines
occurring on histones increases to 84% (Fig. 4d). These estimates
do not account for several sites on H2B (K11, K12, K15, K16, and
K23), as well as sites on H2A (K4, K5, K7, K9, K11, and K13).
Thus, histone acetylation likely accounts for an even greater
proportion of the acetylated lysine residues found in cells.
Regulation by deacetylases. We analyzed the stoichiometry of
lysine deacetylase (KDAC)-regulated acetylation sites by com-
paring the data obtained in this study with a previous analysis of
deacetylase inhibitors in HeLa cells21. Class I KDACs (HDAC 1,
2, 3, and 8) are specifically targeted by the class I inhibitors
apicidin, MS-275, valproic acid, and sodium butyrate; the class IIb
KDAC HDAC6 is specifically targeted by tubacin; and nicotina-
mide inhibits the activity of NAD+-dependent Sirtuin deacety-
lases, but mostly affected SIRT1-regulated sites in mammalian
cells21. To analyze class I KDAC regulated sites, we used the
median increased acetylation caused by apicidin, MS-275, val-
proic acid, and sodium butyrate.
KDAC inhibitors regulated sites with a broad range of
stoichiometry (Fig. 5a). Class I KDAC inhibitors regulated
substantially greater proportions of moderately increased
(>0.23%) and high stoichiometry (>1%) acetylation compared
to tubacin or nicotinamide. The stoichiometry of class I regulated
sites was significantly higher than not-regulated (NR) sites, while
the distributions of tubacin and nicotinamide regulated sites was
not significantly different than NR sites. Furthermore, sites that
were most sensitive to KDAC inhibitors (>4× increased
acetylation) showed an increasing proportion of higher stoichio-
metry sites for the class I inhibitors, but stayed the same or
decreased for tubacin and nicotinamide (Fig. 5a). Thus, while
tubacin and nicotinamide regulate a greater portion of the
acetylation sites quantified in this study (8.6% and 8.8%,
respectively) than class I inhibitors (2.6%), the class I inhibitors
regulate a greater proportion of higher stoichiometry acetylation.
Regulation by the CBP and p300 acetyltransferases. The
homologous Creb-binding protein (CBP)/E1A-binding protein
p300 (p300) acetyltransferases are important regulators of cell-
type-specific and signaling-regulated gene expression22. CBP/
p300 acetyltransferase activity is essential for promoting gene
transcription and CBP/p300 targets a large proportion of the
acetylome12. CBP/p300-regulated sites constituted 12.7% of the
sites analyzed in this study, indicating that CBP/p300 targets
more than one out of every ten acetylation sites. CBP/p300 tar-
geted a similar proportion (11.5%) of low stoichiometry ( < 0.2%)
sites, and an increasing proportion of higher stoichiometry sites,
up to 65% of the sites with stoichiometry exceeding 1% (Fig. 5b).
Thus, CBP/p300 acetylates a majority of high (>1%) stoichio-
metry acetylation sites. The stoichiometry of CBP/p300-regulated
sites also increased with the degree of downregulated acetylation
in the absence of CBP/p300 catalytic activity (Fig. 5c), indicating
that the sites most affected by loss of CBP/p300 tend to be more
highly acetylated.
Stoichiometry of functionally characterized sites. The activity
and subcellular localization of eukaryotic translation initiation
factor 5 A (eIF5A) is regulated by PCAF-catalyzed acetylation at
K4723. We found that eIF5A was more than 10% acetylated at
K47, consistent with a regulatory role for acetylation at this
position. DNA methyltransferase 1 (DNMT1) is acetylated by the
TIP60 acetyltransferase, and acetylation promotes ubiquitin-
0.001%
0.01%
0.1%
1%
10%
100%
Stoichiometry
0.001%
0.01%
0.1%
1%
10%
100%
Stoichiometry
2–4x reduced
4–8x reduced
>8x reduced
Unregulated
>2x
>4x
NR
>2x
>4x
NR
>2x
>4x
NR
Class I KDAC
inhibitors
Tubacin Nicotinamide
2.1%
16.7%
20.0%
3.6%
2.7%
2.7%
2.7%
5.6%
2.7%
9.3%
38.1%
40.0%
13.2%
8.9%
9.1%
10.5%
12.7%
8.1%
Fraction >1%
Fraction >0.23%
P = 2.5e–7 N.S. N.S.
0%
10%
20%
30%
40%
50%
60%
70%
All sites
<0.2%
>0.2%
>0.5%
>1%
Fraction regulated by CBP/p300
Stoichiometry
abc
Fig. 5 Stoichiometry of deacetylase- and CBP/p300-regulated acetylation sites. aThe category scatterplot shows the distributions of acetylation sites that
are not regulated (NR), more than two-fold (>2×) upregulated, or more than four-fold (>4×) upregulated, by the indicated deacetylase inhibitors as
determined by21. Class I KDAC inhibitors primarily target HDACs 1, 2, 3, and 8, and were determined by the median SILAC ratio of apicidin, MS-275,
valproic acid, and sodium butyrate-treated HeLa cells. Tubacin is an HDAC6 inhibitor and nicotinamide inhibits Sirtuin deacetylases, but the regulated sites
are mostly attributed to SIRT121.bCBP/p300 regulates an increasing fraction of high stoichiometry acetylation sites. CBP/p300-regulated sites were
determined by12.cAcetylation sites that are most affected (>8× reduced) by loss of CBP/p300 activity have higher median stoichiometry than sites that
are only modestly affected (2–4× reduced). Source data are provided as a Source Data file
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0 ARTICLE
NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved
dependent protein turnover24. We found that DNMT1 harbored
high stoichiometry acetylation at K335 (0.57%) and K675 (0.1%).
Pyruvate dehydrogenase E1 alpha 1 subunit (PDHA1) is acety-
lated at K321 by acetyl-CoA acetyltransferase 1 (ACAT1), which
recruits PDH kinase (PDK) to inhibit pyruvate dehydrogenase
activity25. We found that PDHA K321 was 0.6% acetylated,
supporting high stoichiometry acetylation at this position.
Glucose-6-phosphate dehydrogenase (G6PD) is reported to be
negatively regulated by KAT9-dependent acetylation of K403, and
activated by SIRT1-dependent deacetylation26. We found that
G6PD is just 0.02% acetylated at K403. Acetylation of
phosphoglycerate kinase 1 (PGK1) at K220 disrupts its activity
by inhibiting binding to ADP27. We found that PGK1 is 0.03%
acetylated at K220 in HeLa cells. PCAF is reported to regulate
cyclin dependent kinase 2 (CDK2) activity by acetylating CDK2
at K33 in the ATP-binding active site28. We found that CDK2
K33 is just 0.05% acetylated, however, CDK1 K33 was 4.5%
acetylated, suggesting that acetylation may play a greater role in
regulating CDK1 activity3. In each of the above examples,
acetylation is reported to reduce enzymatic activity, however, the
stoichiometry of acetylation suggests that acetylation would need
to be dramatically increased (100-fold or more) in order to have a
substantial impact on protein function.
Low stoichiometry acetylation does not necessarily indicate a
lack of function. Rather, the mechanism-of-action determines
whether low stoichiometry acetylation is sufficient to regulate
protein function. Acetylation that imparts a gain-of-function or
regulates protein activity at a specific time and/or place, could be
regulated by low stoichiometry acetylation. For example, acetyla-
tion at K220 regulates the activity of microtubule associated
protein RP/EB family member 1 (EB1), specifically during mitosis
and at spindle microtubule plus ends29. Although we found that
EB1 K220 was only 0.02% acetylated, the mechanism-of-action is
consistent with the low observed stoichiometry of modification.
Discussion
Here we provide a validated resource of acetylation stoichiometry
at thousands of sites in the most widely studied human cell line
(HeLa). Our data shows that the vast majority of acetylation
occurs at very low stoichiometry. Thus, as a general rule, the
mechanisms by which acetylation regulates protein function
should agree with a low stoichiometry of modification. There are
a large number of metabolic enzymes whose catalytic activity is
reduced by site-specific acetylation30. Many studies relied on
acetylation mimicking glutamine-substitution mutations to assay
the impact of acetylation on these proteins, a method that results
in 100% stoichiometry of modification. However, our measure-
ments indicate that the vast majority of acetylation occurs at a
stoichiometry that is much less than 1%, and is therefore not
likely to impact protein activity through a loss-of-function
mechanism at a single acetylation site. Low stoichiometry acet-
ylation is compatible with gain-of-function mechanisms, or in
processes that are spatially or temporally restricted. However,
even for gain-of-function mechanisms such as increased catalytic
activity, a high degree of acetylation may be required to have a
measurable impact on overall activity. Thus, understanding the
stoichiometry of acetylation is important for formulating accurate
mechanistic models when evaluating the impact of acetylation on
protein function. High stoichiometry sites may be particularly
interesting because they are more likely to be enzyme-catalyzed.
However, more studies are required to determine if high stoi-
chiometry is a good indicator of functional importance. Enzy-
matic acetylation does not necessarily indicate a regulatory
function, and high stoichiometry acetylation may also occur by
nonenzymatic mechanisms.
We previously showed that the Sirtuin deacetylases SIRT3 and
CobB suppress acetylation at hundreds of sites to levels that are
equal to or less than the median stoichiometry of modification7,8.
These data support the idea that Sirtuin deacetylases may have a
general role suppressing nonenzymatic acetylation to preserve
protein function31. Here we find that SIRT1 and HDAC6 also
suppress acetylation at regulated sites to levels that are compar-
able to the median stoichiometry of acetylation. SIRT1 likely
suppresses the activity of nuclear acetyltransferases, ensuring
tight control over the sites targeted by these enzymes. HDAC6
deacetylates a large number of cytoplasmic acetylation sites, most
of which are acetylated by unidentified acetyltransferases, or
nonenzymatically. Interestingly, the HDAC6 inhibitor Bufexamac
sensitizes cells to nonenzymatic acetylation caused by aspirin32,
suggesting that HDAC6 may have a role in protecting cells from
nonenzymatic acetylation stress. It is important to note that the
activity of these deacetylases as general suppressors of acetylation
at hundreds of sites, or as dynamic regulators of individual
acetylation sites, is not mutually exclusive. Deacetylases may also
target hundreds of acetylation sites for no reason whatsoever; if
such activity is not evolutionarily disadvantageous, it will not be
selected against.
Our data highlights the outsized impact of CBP and p300 on
the acetylome. CBP/p300 targets up to 20% of all acetylation sites
in cells, and acetylates proximal proteins in a sequence-
independent manner12. Here we find that CBP/p300 targets a
majority (65%) of high (>1%) stoichiometry acetylation. How-
ever, CBP/p300 also acetylated many sites to a low stoichiometry
of modification. These low stoichiometry sites likely include both
functionally important regulatory acetylation and non-functional,
off-target acetylation. This presents a challenge to prioritizing
sites for mechanistic analyses and suggests that additional para-
meters, such as dynamic turnover rates12 and conditional reg-
ulation, are needed to identify potentially interesting sites.
Several other groups have also used chemical acetylation to
measure acetylation stoichiometry, often reporting substantially
higher stoichiometry than we observed in our experiments19,33,34.
One crucial difference between these studies and our own is that
we use a low degree of chemical acetylation (~5–10%) and dilute
the fraction of chemically acetylated peptides to allow for accurate
quantification8. Other studies used stable-isotope-labeled acet-
ylating agents, such acetic anhydride and NHS-acetate, to com-
pletely acetylate all free lysine residues and isotopically label the
chemically acetylated fraction at the same time19,33,34. While this
is an elegant approach, incomplete isotopic labeling of the acet-
ylating agents limits the resolution to stoichiometry greater than
1–2%, and it is not possible to independently dilute the chemi-
cally acetylated peptides to ensure accurate quantification. Most
of these studies19,33–35 did not validate their results using
orthogonal methods, and the accuracy of their measurements is
likely impacted by the limited dynamic range of accurate quan-
tification by mass spectrometry36,37. One study found that ~75%
of their measurements were impacted by false quantification19,
and we found that ~90% of our measurements were inaccurate
when quantifying the differences between completely (100%)
acetylated peptides and native acetylated peptides8. The differ-
ences between 100% acetylated and native acetylated peptides are
likely too large to be accurately quantified by MS. Given the
differences in acetylation stoichiometry reported by independent
studies, care should be taken to carefully validate these
measurements.
The paucity of high stoichiometry acetylation detected in our
study is somewhat disappointing. However, as we show, it is
consistent with the inability to detect acetylated peptides in deep
proteome measurements. One possibility is that proteins are
deacetylated during protein extraction, resulting in uniformly
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0
8NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
decreased acetylation. We consider this unlikely for several rea-
sons. We and others can quantify acetylation changes that occur
in vivo, indicating that these differences are preserved during
protein extraction. We furthermore compared acetylation after
lysing cells by three different methods and found that acetylation
levels were unaffected by the method of protein extraction
(Supplementary Figure 3). Thus, we believe that our measure-
ments indicate the stoichiometry of acetylation that occurs on
proteins inside cells.
Methods
Cell culture. HeLa (ATCC: CCL-2) cells were tested for mycoplasma con-
tamination and grown in DMEM supplemented with 10% FBS, 2 mM L-glutamine,
and 1% penicillin/streptomycin. SILAC media was supplemented with arginine and
lysine (SILAC Light) or with heavy isotope-labeled arginine (13C
6,
15N
4
-arginine,
Sigma) and lysine (13C
6,
15N
2
-lysine, Cambridge Isotope Laboratories) (SILAC
heavy) in media containing dialyzed serum (Sigma). Cells were cultured at 37 °C in
a humidified incubator at 5% CO
2
. At a confluency of ~90%, cells were washed
twice with PBS and lysed in ice-cold modified RIPA buffer (50 mM Tris, pH 7.5,
150 mM NaCl, 1 mM EDTA, 1x mini complete protease inhibitor cocktail (Roche),
10 mM nicotinamide, and 5 μM trichostatin A). Lysates were mixed with 1/10
volume of 5 M NaCl to release chromatin-bound proteins and incubated for 15 min
on ice. Subsequently, lysates were homogenized by sonication (6 × 10 sec, 15 W),
cleared by centrifugation (20,000 × g, 15 min, 4 °C), and the supernatant pre-
cipitated by addition of four volumes of −20 °C acetone. Precip itates were re-
dissolved in 8 M guanidine HCl, 50 mM Hepes pH8.5 and protein concentration
was determined by Quick-start Bradford assay (Bio-Rad).
Chemical acetylation. Protein lysates in 8 M guanidine HCl were mixed with 1/10
volume 1 M acetyl-phosphate (Sigma) prepared in H
2
O, and the acetylation
reaction was allowed to proceed for 2 h at 37 °C. Control reactions were prepared
by mixing with 1/10 volume H
2
O. The acetyl-phosphate was quenched by diluting
the reaction five-fold in 8 M Guanidine HCl, 100 mM Tris pH8. Control and
chemically acetylated SILAC-labeled protein lysates were mixed in equal portions
according to the protein concentration as determined by Bradford assay (above).
The samples were digested (described below) and analyzed by mass spectrometry
to determine the degree of partial chemical acetylation (PCA) (Supplementary
Figure 1a). The measured SILAC ratios were additionally used to adjust the mixing
of SILAC lysates in subsequent experiments.
Protein digestion. Protein lysates were reduced and alkalated with 5 mM TCEP
and 5 mM chloracetamide for 45 min at room temperature. Approximately 10 mg
protein per condition was diluted to 2 M guanidine HCl with 50 mM Hepes, pH8
and digested by endoproteinase Lys C (1:200 w/w; Wako) for 2 h at room tem-
perature. The lysates were further diluted to 1 M guanidine HCl and digested with
trypsin protease (1:200 w/w; Sigma-Aldrich) for 16 h at 37 °C. Digestion was
stopped by the addition of trifluoroacetic acid (TFA) to a final concentration of 1%.
Digests were cleared by centrifugation (2500 × g, 5 min) and loaded onto reversed-
phase C18 Sep-Pak columns (Waters), pre-equilibrated with 5 ml acetonitrile and
2 × 5 ml 0.1% TFA. Peptides were washed with 0.1% TFA and H
2
O, and eluted
with 50% acetonitrile.
Acetylated peptide enrichment. Peptide concentration was determined by UV
spectrometry. The first experiment was performed in technical replicates using two
different peptide fractionation strategies. In the first strategy, peptides were pre-
fractionated by high pH reversed-phase chromatography11 to 6 fractions followed
by acetylated peptide enrichment and micro-scale (in stage-tip) scale strong cation
exchange (SCX) chromatography to an additional 3 fractions (pH4.5, 5.5, and 9.0),
for a total of 18 fractions. The second strategy was to perform acetylated peptide
enrichment on the entire peptide digest, followed by micro-scale SCX into 5
fractions (pH3.5, 4.0, 4.5, 5.5, and 9.0). The second experiment was performed
using the second fractionation strategy only. For acetylated peptide enrichment, the
peptides were mixed with 100 µl 10x IP buffer (500 mM MOPS; pH 7.2, 100 mM
Na-phosphate, 500 mM NaCl, 5% NP-40) per 5 mg peptides. The acetonitrile was
removed and the volume reduced to ~1 ml, by vacuum centrifugation. The final
volume was adjusted with H
2
O to a concentration of 5 mg/ml. 40 μl of anti-
acetylated lysine antibody (PTMScan Acetyl-Lysine Motif [Ac-K] Kit, Cell Sig-
naling Technology) was washed 3× in 1 mL IP buffer, the peptides were clarified by
centrifugation at 20,000 × gfor 5 min, and the peptide supernatant was mixed with
the anti-acetylated lysine antibody. Peptides were enriched overnight at 4 °C,
washed 3 × in 1 ml cold (4 °C) IP buffer, 4× in 1 ml cold IP buffer without NP-40,
and 1× in 1 ml H
2
O. All wash buffer was removed using a 26 gauge needle on an
aspirator. Acetylated peptides were eluted with 100 μl 0.15% TFA, repeated for a
total of three times. Peptides were loaded directly onto a micro-SCX column,
fractionated as described above, and de-salted on C18 stage-tips38.
Mass spectrometry. Peptides were analyzed by nanoflow liquid chromatography-
coupled tandem mass spectrometry (nLC-MS/MS) using a Proxeon easy nLC 1200
connected to a Q-Exactive HF mass spectrometer (Thermo Scientific). The Q-
Exactive was operated in profile mode using positive polarity and a Top10 data
dependent acquisition (DDA) method with the following settings: Spray voltage =
2 kV, S-lens RF level =50, heated capillary =275 °C. Full scan (MS1) was per-
formed with an m/z range of 300–1750 at 60,000 resolution with a target value of
3×10
6ions and a maximum fill time of 20 milliseconds (ms). Fragment (MS2)
scans were performed at a resolution of 15,000 with a target value of 5 × 104ions, a
maximum injection time of 110 ms, an isolation width of 1.3 m/z, a normalized
collision energy (NCE) of 28, and a fixed first mass of 100 m/z. Peptide were
separated by nanoflow chromatography using an EASY-nLC 1000 system (Thermo
Scientific) connected to a 15 cm capillary column packed with 1.9 μm Reprosil-Pur
C18 beads (Dr. Maisch). Column temperat ure was maintained at 40 °C using an
integrated column oven (PRSO-V1, Sonation GmbH) Peptides were eluted by a
gradient of acetonitrile (ACN) in 0.1% formic acid. A typical run utilized a 120 min
gradient followed by 15 min wash and equilibration. A linear gradient at 250 nl/
min of 8-24% ACN (90 min) and 24–40% ACN (30 min) was followed by a wash at
40–80% ACN (5 min) and 80%–8% ACN (5 min), and equilibration at 8% ACN (5
min).
Raw MS data were analyzed using MaxQuant (developer version 1.5.5.4) with
the integrated Andromeda search engine39 to search the UniProt human FASTA
(downloaded 6 July 2015). The following Andromeda settings were used; initial
search mass tolerance of 20 ppm, main search mass tolerance of 6 ppm for parent
ions and 20 ppm for HCD fragment ions, trypsin specificity with a maximum of
two missed cleavages, cysteine carbamidomethylation as a fixed modification, and
N-acetylation of proteins, oxidized methionine, and acetylated lysine as variable
modifications. Acetylated peptides were filtered for a minimum Andromeda score
of 40, as per the default settings for modified peptides. Known contaminants were
removed based on classification by MaxQuant. The false discovery rate (FDR) was
estimated for peptides and proteins individually using a target-decoy approach
allowing a maximum of 1% false identifications from a reversed sequence database.
Calculation of acetylation stoichiometry. We used MaxQuant to analyze the
SILAC ratios of native and chemically acetylated peptides. In order to accurately
calculate stoichiometry it is important to compare the SILAC ratios of individual,
singly acetylated peptides from the “evidence.txt”table. We do not use the “Acetyl
(K)Sites.txt”table since the SILAC ratios of individual sites are sometimes derived
from multiple peptides. Likewise, entries in the “modificationSpecificPeptides.txt”
table can include different positions of acetylation within the same peptide
sequence. Starting with the “evidence.txt”table the following actions were
performed;
1. Remove all entries where Reverse =+, Potential contaminant =+, Acetyl
(K) =0, and Acetyl (K) =2. This removes reverse and contaminant entries
and results in only entries containing singly acetylated peptides.
2. The Modified Sequence was used as a unique identifier to calculate the median
SILAC ratio and summed peptide intensity for multiple instances of any given
acetylated peptide in each experiment.
3. SILAC ratios were tested for agreement with the dilution series, allowing for
up to two-fold variability8. Stoichiometry was calculated using only peptides
that met these strict criteria.
4. Stoichiometry was calculated as follows; Stoichiometry (S), degree partial
chemical acetylation (C), dilution factor for acetylated peptides (D), and
SILAC ratio partial chemical acetylation/native acetylation (R). S =(C)/
((R*D)-(1-C)). The dilution factors (D) were as follows: Experiment 1, (~1%
D=4.23), (~0.1% D =42.3), (~0.01% D =423), Experiment 2, (~1% D =
6.37), (~0.1% D =63.7), (~0.01% D =637). The median degree of partial
chemical acetylation (C) was 3.53% and 10.38%, for experiments 1 and 2,
respectively, (Supplementary Figure 1a).
The above mentioned .txt files can be found via the PRIDE40 partner repository
with the dataset identifier PXD009994. Relative acetylation stoichiometry was
estimated using acetylation site intensity corrected for differences in protein
abundance, otherwise referred to as abundance-corrected intensity (ACI). ACI was
calculated by dividing acetylation site intensity by iBAQ protein abundance.
Validation by AQUA and spike-in. Two different unmodified AQUA peptides
(Thermo Scientific, AQUA quant pro) per protein were added to SILAC heavy-
labeled HeLa peptides to determine the concentration of each protein (CTTN,
NCL, and NAT10). Acetylated AQUA peptides were then added to final stoi-
chiometry of 1%, 0.1%, or 0.01%, in accordance with each individual protein’s
concentration. Acetylated peptides were enriched as described above and analyzed
by mass spectrometry. For validation by recombinant acetylated protein spike-in,
recombinant acetylated protein was added to HeLa lysate and digested to deter-
mine the concentration of protein present in the lysate. Based on the initial ana-
lysis, recombinant HMGCL K48ac was added at a 1:1 (100%), 1:10 (10%), and
1:100 (1%) stoichiometry and MDH2 K239ac was added at a 1:1 (100%), 1:1,000
(0.1%), and 1:10,000 (0.01%) stoichiometry. Acetylated peptides were enriched
from the 1:10 and 1:100 dilutions (HMGCL) and the 1:1,000 and 1:10,000 dilutions
(MDH2) before analysis by mass spectrometry. The 1:1 dilutions were analyzed to
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0 ARTICLE
NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved
determine the protein mixing ratio and to correct the stoichiometry measurements
accordingly. The mixing ratio of spike-in/HeLa was 1.48 for HMGCL and 0.70 for
MDH2.
Expression and purification of recombinant acetylated proteins. MDH2
(K239AcK) and HMGCL (K48AcK) were expressed from a pRSF-Duet-vector
containing the HMGCL coding region with an amber-stop codon at the position
encoding for acetyl-L-lysine incorporation, and additionally the coding regions for
a synthetically evolved pair of Methanosarcina barkeri MS tRNA
CUA
(MbtRNA-
CUA) and the acetyl-lysyl-tRNA-sythetase (AcKRS3). AcKRS3 and MbtRNA
CUA
enable the site-specific incorporation of acetyl-L-lysine into proteins as response to
an amber codon. This was done by supplementing the E. coli BL21 (DE3) culture
with 10 mM N-(e)-acetyl-lysine (Chem-Impex International, Inc.) and 20 mM
nicotinamide to inhibit the E. coli CobB deacetylase at an OD
600
of 0.6 (37 °C, 160
rpm). Subsequently, the temperat ure was lowered to 18 °C and cells were grown for
another 30 min at 160 rpm. Protein expression was induced by addition of 200 µM
IPTG and protein expression conducted overnight (18 °C, 160 rpm). The cells were
harvested by centrifugation (4000 × g, 20 min) and resuspended in buffer A. Cell
lysis and protein purification was performed as described above.
Histone western blot. HeLa cells were lysed by boiling in 2% lithium dodecyl
sulfate (LDS) (1x NuPAGE LDS Sample Buffer, ThermoFischer Scientific), fol-
lowed by sonication to disrupt genomic DNA. Proteins were separated on a 4-12
NuPAGE gel (ThermoFischer Scientific) and transferred to a nitrocellulose
membrane (BioRad). Acetylated proteins were visualized by immunoblot using
pan-anti-acetylated lysine antibody (Immunechem #ICP0380) at a 1/1000 dilution
and goat anti-rabbit horseradisch peroxidase (HRP) (BioRad, STAR124P) sec-
ondary antibody at a 1/5000 dilution.
Data analysis. Pearson’s correlation (r) and Wilcoxon tests were performed in R.
UniProt keyword analysis was performed using AGOTOOL10 with an uncorrected
P-value cutoff of 0.05. Word clouds were generated in R. To estimate histone
acetylation copy numbers we used the copy number estimate for histone H4 (5e7
per cell) for all four core histones. This is because copy number estimates for
histones are complicated by the presence of multiple isoforms of H2B, H2A, and
H3. The copy number estimate for H4 is consistent with a previous analysis that
estimated that core histones represent ~4% of the total protein in HeLa cells, and
with the DNA content of HeLa cells41. We furthermore used the median stoi-
chiometry for sites that occurred on multiple histone isoforms (such as H2B K5) to
avoid over-counting these sites.
Reporting summary. Further information on experimental design is available in
the Nature Research Reporting Summary linked to this article.
Data availability
The raw mass spectrometry data have been deposited to the ProteomeXchange
Consortium via the PRIDE40 partner repository with the dataset identifier PXD009994.
The source data underlying Figs. 1a, 2a-b, 2d-e, 3a-d, 4a, 4c-d, 5a-c and Supplementary
Figs. 1a-c, 2a-c, and 3 are provided as a Source Data file. A reporting summary for this
Article is available as a Supplementary Information file. All other data supporting the
findings of this study are available from the corresponding authors upon reasonable
request.
Received: 20 November 2018 Accepted: 11 February 2019
References
1. Allfrey, V. G., Faulkner, R. & Mirsky, A. E. Acetylation and methylation of
histones and their possible role in the regulation of RNA synthesis. Proc. Natl
Acad. Sci. USA 51, 786–794 (1964).
2. Kim, S. C. et al. Substrate and functional diversity of lysine acetylation
revealed by a proteomics survey. Mol. Cell 23, 607–618 (2006).
3. Choudhary, C. et al. Lysine acetylation targets protein complexes and co-
regulates major cellular functions. Science 325, 834–840 (2009).
4. Lundby, A. et al. Proteomic analysis of lysine acetylation sites in rat tissues
reveals organ specificity and subcellular patterns. Cell Rep. 2, 419–431 (2012).
5. Wagner, G. R., Payne, R. M. Widespread & Enzyme-independent, N {epsilon}-
Acetylation and N{epsilon}-succinylation of proteins in the chemical
conditions of the mitochondrial matrix. J. Biol. Chem. 288, 29036–29045
(2013).
6. Weinert, B. T. et al. Acetylation dynamics and stoichiometry in
Saccharomyces cerevisiae. Mol. Syst. Biol. 10, 716 (2014).
7. Weinert, B. T., Moustafa, T., Iesmantavicius, V., Zechner, R. & Choudhary, C.
Analysis of acetylation stoichiometry suggests that SIRT3 repairs
nonenzymatic acetylation lesions. EMBO J. 34, 2620–2632 (2015).
8. Weinert, B. T. et al. Accurate quantification of site-specific acetylation
stoichiometry reveals the impact of sirtuin deacetylase CobB on the E. coli
acetylome. Mol. Cell. Proteom. 16, 759–769 (2017).
9. Carabetta V. J., Greco T. M., Tanner A. W., Cristea I. M., Dubnau D.
Temporal regulation of the Bacillus subtilis acetylome and evidence for a role
of MreB acetylation in cell wall growth. mSystems 1, e00005–16 (2016).
10. Scholz, C. et al. Avoiding abundance bias in the functional annotation of
posttranslationally modified proteins. Nat. Methods 12, 1003–1004 (2015).
11. Svinkina, T. et al. Deep, quantitative coverage of the lysine acetylome using
novel anti-acetyl-lysine antibodies and an optimized proteomic workflow.
Mol. Cell. Proteom. 14, 2429–2440 (2015).
12. Weinert, B. T. et al. Time-resolved analysis reveals rapid dynamics and broad
scope of the CBP/p300 acetylome. Cell 174, 231–244 e212 (2018).
13. Bekker-Jensen, D. B. et al. An optimized shotgun strategy for the rapid
generation of comprehensive human proteomes. Cell Syst 4, 587–599 (2017).
14. Thul P. J., et al. A subcellular map of the human proteome. Science
356, eaal3321 (2017).
15. Colaert, N., Helsens, K., Martens, L., Vandekerckhove, J. & Gevaert, K.
Improved visualization of protein consensus sequences by iceLogo. Nat.
Methods 6, 786–787 (2009).
16. James, A. M., Smith, A. C., Smith, C. L., Robinson, A. J. & Murphy, M. P.
Proximal cysteines that enhance lysine N-acetylation of cytosolic proteins in
mice are less conserved in longer-living species. Cell Rep. 24, 1445–1455
(2018).
17. Abshiru, N. et al. Discovery of protein acetylation patterns by deconvolution
of peptide isomer mass spectra. Nat. Commun. 6, 8648 (2015).
18. Feller, C., Forne, I., Imhof, A. & Becker, P. B. Global and specific responses of
the histone acetylome to systematic perturbation. Mol. Cell 57, 559–571
(2015).
19. Zhou, T., Chung, Y. H., Chen, J. & Chen, Y. Site-specific identification of
lysine acetylation stoichiometries in mammalian cells. J. Proteome Res. 15,
1103–1113 (2016).
20. Zheng, Y., Thomas, P. M. & Kelleher, N. L. Measurement of acetylation
turnover at distinct lysines in human histones identifies long-lived acetylation
sites. Nat. Commun. 4, 2203 (2013).
21. Scholz, C. et al. Acetylation site specificities of lysine deacetylase inhibitors in
human cells. Nat. Biotechnol. 33, 415–423 (2015).
22. Dancy, B. M. & Cole, P. A. Protein lysine acetylation by p300/CBP. Chem. Rev.
115, 2419–2452 (2015).
23. Ishfaq, M. et al. Acetylation regulates subcellular localization of eukaryotic
translation initiation factor 5A (eIF5A). FEBS Lett. 586, 3236–3241 (2012).
24. Du, Z. et al. DNMT1 stability is regulated by proteins coordinating
deubiquitination and acetylation-driven ubiquitination. Sci. Signal. 3, ra80
(2010).
25. Fan, J. et al. Tyr phosphorylation of PDP1 toggles recruitment between
ACAT1 and SIRT3 to regulate the pyruvate dehydrogenase complex. Mol. Cell
53, 534–548 (2014).
26. Wang, Y. P. et al. Regulation of G6PD acetylation by SIRT2 and KAT9
modulates NADPH homeostasis and cell survival during oxidative stress.
EMBO J. 33, 1304–1320 (2014).
27. Wang, S. et al. Insulin and mTOR pathway regulate HDAC3-mediated
deacetylation and activation of PGK1. PLoS Biol. 13, e1002243 (2015).
28. Mateo, F. et al. The transcriptional co-activator PCAF regulates cdk2 activity.
Nucleic Acids Res. 37, 7072–7084 (2009).
29. Xia, P. et al. EB1 acetylation by P300/CBP-associated factor (PCAF) ensures
accurate kinetochore-microtubule interactions in mitosis. Proc. Natl Acad. Sci.
USA 109, 16564–16569 (2012).
30. Lin, H., Su, X. & He, B. Protein lysine acylation and cysteine succination
by intermediates of energy metabolism. Acs. Chem. Biol. 7, 947–960
(2012).
31. Wagner, G. R. & Hirschey, M. D. Nonenzymatic protein acylation as a carbon
stress regulated by sirtuin deacylases. Mol. Cell 54,5–16 (2014).
32. Tatham, M. H. et al. A Proteomic approach to analyze the aspirin-mediated
lysine acetylome. Mol. Cell. Proteom. 16, 310–326 (2017).
33. Baeza, J. et al. Stoichiometry of site-specific lysine acetylation in an entire
proteome. J. Biol. Chem. 289, 21326–21338 (2014).
34. Gil, J. et al. Lysine acetylation stoichiometry and proteomics analyses reveal
pathways regulated by sirtuin 1 in human cells. J. Biol. Chem. 292,
18129–18144 (2017).
35. Meyer, J. G. et al. Quantification of lysine acetylation and succinylation
stoichiometry in proteins using mass spectrometric data-independent
acquisitions (SWATH). J. Am. Soc. Mass. Spectrom. 27, 1758–1771 (2016).
36. Bakalarski, C. E. et al. The impact of peptide abundance and dynamic range
on stable-isotope-based quantitative proteomic analyses. J. Proteome Res. 7,
4756–4765 (2008).
37. Lau, H. T., Suh, H. W., Golkowski, M. & Ong, S. E. Comparing SILAC- and
stable isotope dimethyl-labeling approaches for quantitative proteomics. J.
Proteome Res. 13, 4164–4174 (2014).
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0
10 NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
38. Rappsilber, J., Ishihama, Y. & Mann, M. Stop and go extraction tips for
matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS
sample pretreatment in proteomics. Anal. Chem. 75, 663–670 (2003).
39. Cox, J. et al. Andromeda: a peptide search engine integrated into the
MaxQuant environment. J. Proteome Res. 10, 1794–1805 (2011).
40. Vizcaino, J. A. et al. The PRoteomics IDEntifications (PRIDE) database and
associated tools: status in 2013. Nucleic Acids Res. 41, D1063–D1069 (2013).
41. Wisniewski, J. R., Hein, M. Y., Cox, J. & Mann, M. A. “proteomic ruler”for
protein copy number and concentration estimation without spike-in
standards. Mol. Cell. Proteom. 13, 3497–3506 (2014).
42. Kulak, N. A., Pichler, G., Paron, I., Nagaraj, N. & Mann, M. Minimal,
encapsulated proteomic-sample processing applied to copy-number
estimation in eukaryotic cells. Nat. Methods 11, 319–324 (2014).
Acknowledgements
We thank the members of our laboratory for their helpful discussions. L.B. and M.L. were
funded by the German Research Foundation (CECAD and DFG Research Grant Number
LA2984/5-1). C.C. is supported by the Hallas Møller Investigator award (NNF14OC0008541)
from the Novo Nordisk. B.T.W. is supported by a grant from the Novo Nordisk Foundation
(NNF15OC0017774). The Novo Nordisk Foundation Center for Protein Research is sup-
ported financially by the Novo Nordisk Foundation (Grant agreement NNF14CC0001).
Author contributions
B.T.W. and C.C. designed the research and supervised the project. B.K.H. performed
stoichiometry measurements and contributed to data interpretation. R.G. performed
validation experiments. L.B and M.L. provided recombinant acetylated proteins. D.L. and
T.N. provided bioinformatics support. B.T.W. analyzed data, prepared figures, and wrote
the manuscript. All authors read and commented on the manuscript.
Additional information
Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467-
019-09024-0.
Competing interests: The authors declare no competing interests.
Reprints and permission information is available online at http://npg.nature.com/
reprintsandpermissions/
Journal peer review information:Nature Communications thanks Xiao-Lu Zhao and the
other anonymous reviewers for their contribution to the peer review of this work. Peer
reviewer reports are available.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2019
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09024-0 ARTICLE
NATURE COMMUNICATIONS | (2019) 10:1055 | https://doi.org/10.1038/s41467-019-09024-0 | www.nature.com/naturecommunications 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
Available via license: CC BY
Content may be subject to copyright.