Nucleic Acids Research, Vol. 19, No. 18
4937-4941
Determination of the DNA-interacting region of the
archaebacterial chromosomal protein MC1. Photocrosslinks
with 5-bromouracil-substituted DNA
Mehmaz Katouzian-Safadi, Bernard Laine1, Francois Chartier1, Jean-Yves Cremet,
Denise Belaiche1, Pierre Sautiere1 and Michel Charlier
Centre de Biophysique Mol§culaire, CNRS, 1A Avenue de la Recherche Scientifique, 45071
Orleans-Cedex 2 and 1URA 409 du CNRS, Institut de Recherche sur le Cancer de Lille,
Place de Verdun, 59045 Lille, France
Received June 11, 1991; Revised and Accepted August 28, 1991
ABSTRACT
Protein MC1 is the major chromosomal protein in
methanosarcinaceae. Using photochemical crosslinking on 5-bromouracil-substituted DNA, we identified
the region of the protein that interacts with it. This
region is located in the C-terminal part of the polypeptlde chain, and the crosslinked amino-aclds are in
the region 74 - 86. Tryptophan 74 is one of the aminoacids crosslinked to DNA.
This method has been successfully used for various proteins:
histones (12), lac repressor (13,14), £coRl (15) and cyclic AMP
receptor protein (16).
We have determined the region of the MCI protein which is
cross-linked to DNA by sequence analysis of the peptides
remaining bound to DNA after proteolytic cleavage.
MATERIAL AND METHODS
INTRODUCTION
In eubacteria, and in two groups of archaebacteria (thermophilic
sulfur-dependent, and methanogen bacteria), small and abundant
DNA-associated proteins have been shown to modify DNA
structure and to be involved in DNA replication or transcription
(see reviews I - 3 ) . Protein MCI, a basic polypeptide of 11 KDa,
is the major protein of the deoxyribonucleoprotein complex of
methanosarcinaceae, a methanogen family (4—6). Experiments
performed with protein MC 1, isolated from Methanosarcina sp
CHT155, show that protein MC 1 protects DNA against thermal
denaturation (5) and induces DNA bending and supercoiling (7).
Through comparative studies of the primary structures of the
protein MCI isolated from four species, and the use of predictive
methods of secondary structure determination, two regions can
be defined in this protein (8-11). The first region contains two
well-conserved sequences (residues 17—35 and 45—58) with high
probability of beta-sheet and alpha-helix conformations separated
by beta-turns. The second region, located in the carboxy-terminal
third of the protein, is highly variable but contains four strictly
conserved proline residues, three of them having a high
probability to form beta-turns.
Our goal in this paper was to identify the region of the protein
interacting with DNA, using a photochemical cross-linking
technique. We used DNA wherein the thymine residues have been
partially substituted by 5-bromouracil (BrUra). The advantage
of BrUra-substituted DNA is that it allows irradiation at
wavelengths long enough (X>300 nm) to reduce significantly
the direct photolysis of both DNA and proteins, with a crosslinking efficiency considerably greater than with normal DNA.
Endoproteinase Glu-C (EC 3.4.21.19) and endoproteinase LysC (EC 3.4.21.50) were purchased from Boehringer. Acetonitrile
for HPLC was from Carlo Erba. Reagents, solvents and
polyethylenimine used for gas-phase sequencing, were purchased
from Applied Biosystems.
Preparation of BrUra -substituted DNA
BrUra-substituted DNA was purified from an E. coli B/rT~
strain, as described previously (14). The percentage of substitution
was assayed by determination of the buoyant density of the DNA
in a CsCl gradient (17) and found to be 75%. Concentrations were
determined spectrophotometrically, assuming an extinction
coefficient of 13000 M~'cm~' per base pair.
Preparation of protein MCI
Methanosarcina sp CHTJ55 (strain 2902) was grown as described
previously (5).
Protein MCI was prepared as already described (4,7). The
protein concentration was determined spectrophotometrically
assuming an extinction coefficient of 11200 M~'cm~' per
protomer at 280 nm.
Preparation of MC1-DNA complexes
Samples (2.4 ml) were prepared by adding, under gentle stirring,
the protein solution to the DNA solution, both in 10 mM sodium
phosphate, pH 7.25. Final concentrations were 15 /tM protein
and 225 /tM DNA (as base pairs), leading to a ratio of one
molecule of protein per 15 base pairs. In these conditions, we
verified that all molecules of protein are bound to DNA.
4938 Nucleic Acids Research, Vol. 19, No. 18
When necessary, dissociation of the complex can be achieved
by increasing the salt concentration to 0.5 M KC1, 0.2 M
potassium phosphate, pH 7.25.
In this paper, low ionic strength buffer refers to 10 mM sodium
phosphate, pH 7.25, and high ionic strength buffer to 0.5 M KC1,
0.2 M potassium phosphate pH 7.25.
Irradiation of the complexes
Irradiations were performed using a high pressure mercury lamp
(Osram HbO 200 Watts) equipped with filters isolating radiations
of wavelengths ranging from 300 to 400 run (14). The fluence
rate in the front of the irradiation vessel was 480 Wm~2, and
the total incident fluence during the crosslinking reaction was
1.73 106 Jm- 2 .
Chromatographic separation of free and crosslinked proteins
or peptides
Free and crosslinked proteins or peptides were separated by
molecular sieving chromatography on AC A 34 gel. The column
(25x2 cm) was equilibrated and eluted with the high ionic
strength buffer.
Determination of the non-crosslinked peptides (Fig.2,
pathway 1)
The crosslinked DNA-protein complex, rid of free protein as
previously described, was dialysed against water and concentrated
by ultracentrifugation at 223000 Xg for 5 hours. After elimination
of the supernatant, the pellet was redissolved in 0.5 ml water
and adjusted to 0.1 M ammonium bicarbonate, 0.7 M NaCl, pH
7.8. The protein was then digested with endoproteinase Glu-C,
at 37°C for 18 hours, using an enzyme to substrate ratio of 1/50
(w/w).
The digest was centrifuged at 223000Xg for 5 hours. Free
peptides remaining in the supernatant were fractionated by
reversed phase HPLC on a C18 jiBondapack column (300x3.9
mm, Millipore) eluted by a gradient of acetonitrile in 0.05%
trifluoroacetic acid.
Amino-acid analyses of peptides were performed as previously
described (9).
Determination of the crosslinked peptides (Fig.2, pathway 2)
The irradiated complex solution (2.4 ml) was adjusted to 4.7 M
urea. Proteolysis by endoproteinase Lys-C was performed at
20°C for 10 hours at a final enzyme-to-substrate ratio of 1/25
(w/w).
The digest was then dialysed against the low ionic strength
buffer to remove urea, adjusted to high ionic strength, and
chromatographed on ACA 34. The fractions containing the
crosslinked peptides bearing DNA were pooled. The
nucleopeptide was precipitated with cold ethanol, centrifuged,
dissolved and dialysed against distilled water.
Ammonium bicarbonate was added to 0.2 M, before a second
extensive dialysis against bidistilled water. This step allows the
replacement of all contaminating ions and soluble compounds
by a volatile salt.
After concentration to 0.27 ml by evaporation, the sample was
directly submitted to sequencing.
Automated Edman degradation was carried out on a gas-phase
Applied Biosystems sequencer, using the 03RPE1 program, with
polyethylenimine (Applied Biosystems) as carrier (19).
Phenylthiohydantoin (PTH) derivatives of amino-acids were
identified by reversed phase HPLC with an on-line Applied
Biosystems 120 amino-acid analyser.
RESULTS
Cross-linking of protein MCI to DNA
After irradiation at low ionic strength, the complex solution was
adjusted to high ionic strength, to completely remove noncrosslinked proteins from the DNA. Fig. 1 shows the separation
DNA-MC1 complex
I
Irradiation
Cross-linked material
and frae proteins
PATHWAY 1
ACA 34 chromato.
Relative absorbance or fluorescence.
PATHWAY 2
Endo.Lys-C
digestion
Frit
prottln
MC1 crosalinked to DNA
I
ACA 34 chromato.
0.8 1Endo.Qlu-C
digestion
I
Ultracentrifugation
Pallet
DNA orost*
llnkad to
Supernatant
piptlda(t)
Free peptides
I
Peptide(s) crosslinked to ONA
Sequence
determination
HPLC
6
11
16
21
26
31
Fraction number.
Figure 1. Separation of non-crosslinked and crosslinked protein MCI by molecular
sieving chromatography. 2.4 m] of irradiated complex was loaded on the ACA34
column (20x2cm)andeftitedathigh ionic strength with a flow rate of 12 ml/hour.
Fractions of 1.5 ml were monitored by absorbance at 260 nm and by fluorescence
at 345 nm (excitation at 290 nm).
/
Amino-acids analysis
Identification of
free peptides
Identification of
the bound
peptide(s)
Figure 2. Strategies used to determine the DNA-crosslinked region of protein
MCI.
Nucleic Acids Research, Vol. 19, No. 18 4939
of crosslinked material from the free protein, on an ACA 34
column eluted at high ionic strength. Peak I, eluted in the void
volume, contains the DNA bearing the crosslinked proteins. Peak
II corresponds to free protein. From the relative area of the
fluorescence of each peak, an average value of 20% of
crosslinked proteins was determined.
To ascertain that crosslinks do not result from a diffusion
controlled reaction, a mixture of DNA and protein was irradiated
at high ionic strength which prevents complex formation. No
crosslinking was detected under these conditions.
Strategies used to determine the DNA binding domain of
protein MCI
At the beginning of our work, we intended to identify, after
proteolysis, the peptides crosslinked to DNA and the peptides
not bound to DNA (Fig. 2, padiway 1). For this purpose, the
non- crosslinked proteins were eliminated by molecular sieving
chromatography prior to digestion of the crosslinked material with
endoproteinase Glu-C. This enzyme was chosen to obtain a low
number of peptides (Fig. 3 A). Free peptides could be identified
after their fractionation by HPLC. We did not succeed in
identifying the peptide crosslinked to DNA, since low yields were
obtained at each step of purification.
This led us to use a simpler strategy (Fig. 2, pathway 2). Just
after irradiation, both crosslinked and non-crosslinked proteins
were digested with endoproteinase Lys-C. Free peptides were
eliminated by molecular sieving chromatography, and the
crosslinked peptides were directly identified by microsequencing.
Identification of the non-crosslinked peptides (Fig.2,
pathway 1)
The protein MCI crosslinked to DNA was cleaved with
endoproteinase Glu-C in 0.1 M ammonium bicarbonate (pH 7.8)
220 nm
containing 0.7 M NaCl. The high salt concentration, which
abolishes the electrostatic bonds between the protein and DNA,
was expected to increase the accessibility of potential sites of
cleavage. Moreover, to avoid eventual cleavage of the crosslinks
in acidic medium (20), neutral conditions were preferred for die
digestion by the endoproteinase.
Fig. 3A shows the HPLC elution pattern of the peptides
released from the crosslinked protein. The schematic
representation of these peptides in die protein sequence is shown
in Fig. 3B.
The diree peptides obtained from the crosslinked material were
identified by their amino-acid compositions. Peptides PE1
(residues 1 —63) and PEla (residues 1 -49) overlap the aminoterminal part of the protein. Peptide PE2 (residues 88-93)
overlaps the carboxy- terminus of the protein. No trace of a
peptide overlapping the sequence 64-87 was detected.
These results clearly show that there is no crosslinking site
in the region of die protein corresponding to peptides PE1 and
PE2, and strongly suggest diat the crosslinked residue(s) is(are)
located between amino-acids 64 and 87.
We can notice diat glutamyl bonds at positions 11,15 and 90
were not cleaved by the endoproteinase Glu-C in the crosslinked
protein. In die same conditions of digestion, the glutamyl bonds
at positions 11 and 90 were not cleaved eidier when protein MCI
is not crosslinked. We previously observed a lack of cleavage
of glutamyl bonds in the course of die determination of die
sequence of protein MCI variants from M.soehngenii (10). It
is not however clear at present if die lack of cleavage at position
15, and die partial cleavage at position 49 result from a particular
structure of die crosslinked protein, or from a steric hindrance
due to DNA crosslinking.
Identification of the peptides bound to DNA (Fig. 2,
pathway 2)
Free native MCI protein is cleaved wim die endoproteinase Lys-C
at every potential sites. In contrast, when complexed to DNA,
protein MCI remains resistant to attack by die protease.
Nevertheless, this resistance was completely removed in die
presence of 4.7 M urea. These conditions were used for die
crosslinked protein.
After proteolysis, die peptides crosslinked to DNA were
separated from free peptides by chromatography on die ACA
34 column. The crosslinked DNA-peptide(s), purified as
pMoles of PTH derivatives of amino-aclds
1000/
20
30
Elution volume (ml)
N
RP
•7 0>
-I-
B
100.
E
|lVDAPINBPAWUPEII5tPFVIIE|
64
87
Figure 3. A: Reversed phase HPLC elution pattern of the peptides generated
by endoproteinase Glu-C digestion of protein MCI crosslinked to DNA. Full
line, absorbance at 220 nm (left scale). Dashed line, acetonitrile concentration
(right scale). Flow rate: 1 rnl/min. B: Schematic drawing of the non-crosslinked
peptides generated by the cleavage of protein MCI crosslinked to DNA with
endoproteinase Glu-C. The one-letter sequence of the missing part of the protein
is shown.
1
2
3
4
S
6
7
8
9
10 11 12 13 U
15 14 17 18
Cycle n u m b e r .
Figure 4. Edman degradation of crosslinked peptides (700 pmoles) generated
by cleavage of crosslinked protein MCI with endoproteinase Lys-C. PTH
derivatives of amino-acids were identified and quantified by HPLC.
4940 Nucleic Acids Research, Vol. 19, No. 18
1000
Cumulated pMolet oJ PTH-amlno-aclds
100:
E I V D A P I N E P A W M P E t l S t P F V I I E
05
70
76
80
86
Amlno-acld
Figure 5. Same data as Fig. 4, but the two overlapping peptides have been phased
along the sequence of the protein and the amounts of PTH derivatives have been
cumulated. Arrows indicate the non-identified amino-acids.
described in the Material and Methods section, were submitted
to micro-sequencing.
By comparison with the sequence of protein MCI (8), the
results obtained over 18 cycles of automated Edman degradation
(Fig. 4) show that two peptides were simultaneously sequenced.
The major peptide began at Asp 70 and was sequenced as far
as Glu 87. The minor one began at Glu 63 and was sequenced
as far as Glu 77. The repetitive yield of the major peptide is lower
than that of the minor peptide; however, on the basis of the
amounts of the PTH derivatives obtained in the three first steps
of the Edman degradation, the major peptide is at least 2.5 times
more abundant than the minor one.
The diagram showing the cumulated picomoles of PTH
derivatives of amino-acids identified during the sequence analysis
is presented in Fig. 5.
DISCUSSION
Photochemical crosslinking between a protein and DNA can
occur only when bromouracil and a crosslinkable amino-acid are
present in a favourable conformation in the interaction site. The
presence of a crosslink reveals the proximity of residues, but its
absence gives no information. In other words, crosslinking is
a positive, but not an exclusive result. We used DNA where 75 %
of the thymines were substituted by BrUra, i.e. about one fifth
of the total bases are photoreactive residues. As the protein MCI
does not recognize a specific DNA sequence, the crosslinking
yield is limited by the percentage substitution of the DNA. This
explains the observed value of 20%.
Micro-sequencing results (Fig. 5), together with the
identification of the non-crosslinked peptides (Fig. 3B),
unambiguously show that the region interacting with DNA is
located within residues 6 9 - 8 7 . Sequencing of peptides linked
to DNA shows that two overlapping peptides remain bound: the
peptides 6 3 - 7 6 (minor peptide) and 70 - 87 (major peptide). The
good yields of PTH-amino-acids shows that no crosslink occurred
in the sequence 6 3 - 7 3 (Fig. 5). Therefore, we can infer that
recovery of the minor peptide results from a partial cleavage of
the lysine bond at position 69. This can be explained by its
restricted accessibility due to the steric hindrance of the
crosslinked DNA. Such a restricted accessibility also explains
lack of cleavage at lysine residues 78, 81, 85 and 86.
Lack of identification of Serine 80 could result from its
degradation after eleven cycles of automatic sequencing.
However, the low yield of lysine 78, and the lack of identification
of lysines 81, 85 and 86 cannot be explained at present. Lysine
residues can be crosslinked to DNA, but simultaneous
involvement of these residues in the crosslinking reaction seems
very unlikely, since it would need the presence of BrUra clusters.
There are probably several pathways to crosslink BrUrasubstituted DNA and proteins. Model systems using amino-acids
or peptides have been studied, including cysteine or cystine
(21,22), tryptophan, tyrosine and histidine (23-26), lysine (25)
and even the peptide bond (27). Among these putative
crosslinking sites, tryptophan residue is one of the more reactive,
and can be easily crosslinked to BrUra. In the case of protein
MCI, tryptophan 74 which occupies positions 5 and 12 in the
major and minor peptides respectively, was not identified during
sequencing of crosslinked peptides (Fig. 4). This strongly
suggests the involvement of tryptophan 74 in the crosslinking
reaction. Moreover, when protein MCI is complexed to DNA,
the fluorescence in the tryptophan band decreases by 20%,
without any change in the shape of the spectrum (Culard, personal
communication). Such a result has already been observed with
proteins or peptides, whose tryptophan residues interact directly
with DNA (28,29).
The interacting sequence 69-87 contains one arginine and five
lysine residues, which could give electrostatic interactions with
DNA. It does not correspond to the most basic region of the
protein, which is located in the middle part of the protein (residues
46—55), where half of the residues are arginine or lysine. The
interacting sequence contains 10 highly conserved residues found
in every MCI protein from Methanosarcinaceae studied up to
now (10). Among these residues are the three lysines at positions
81, 85 and 86 and three prolines at position 72, 76 and 82.
According to secondary structure predictions (10), prolines 72
and 76 may form adjacent beta-turns, and thus may play a role
in the conformation of this region (30). The high conservation
of these residues appears to be critical for the DNA binding ability
of protein MC1.
The DNA binding region of protein MCI differs from the
helix-turn-helix motif encountered in several bacterial DNAbinding proteins (31). On the other hand, this region is
reminiscent to the DNA-binding region of protein HU, an
abundant chromosomal protein in eubacteria (1,2). The
B.stearothermophilus protein HU interacts with DNA through
an external arm consisting of beta-sheet strands separated by betatums (32). The hydrophobic clusters analysis method (33) shows
that the interacting region of protein MC 1 displays hydrophobic
clusters with shape and orientation characteristic of beta-sheet
ribbons, flanked by the highly conserved proline residues cited
above (18). This suggests that as protein HU, protein MCI
interacts with DNA through beta-sheet structures separated by
beta-turns.
ACKNOWLEDGEMENTS
The authors are indebted to J.P.Touzel, from the Station de
Technologie Alimentaire (INRA, Villeneuve d'Ascq) for
providing the bacteria. We acknowledge the technical assistance
of A.Lemaire-Poitau. We thank J.Rossier, J-P.Le Caer, and
F.Culard for helpfull discussions, and communication of
preliminary results. This work was partially supported by grants
from the CNRS, the University de Lille II, and the Pdle des
Anairobies de la re'gion Nord-Pas de Calais.
Nucleic Acids Research, Vol. 19, No. 18 4941
REFERENCES
1. Driica, K. and Rouviere-Yaniv, J. (1987) Microbiol. Re\iews, 51, 301 -319.
2. Dijk, J. and Reinhardt, R. (1986) In Guakrzi, C O . and Pon, C.L. (eds.)
Bacterial Chromatin. Springer-Verlag. pp 186-218.
3. Schmid, M.B. (1990) Cell, 63, 461-451.
4. Chartier, F., Laine, B., Sautiere, P., Touzel, J-P. and Albagnac, G. (1985)
FEBS Lett., 183, 119-123.
5. Chartier, F., Laine, B. and Sautiere, P. (1988) Biochim. Biophys. Acta, 951,
149-156.
6. Imbert, M., Laine, B., Prensier, G., Touzel, J-P. and Sautiere, P. (1988)
Can. J. Microbiol.. 34, 931-937.
7. Laine, B., Culard, F., Maurizot, J-C. and Sautiere, P. Nucleic Acids Research,
in press.
8. Chartier, F., Laine, B., Belalche, D., Touzel, J-P. and Sautiere, P. (1989)
Biochim. Biophys. Acta, 1008, 309-314.
9. Laine, B., Chartier, F., Imbert, M., Lewis, R. and Sautiere, P. (1986) Eur.
7. Biochem., 161, 681-687.
10. Chartier, F., Laine, B., Belalche, D. and Sautiere, P. (1989)7. Biol. Chem.,
264, 17006-17015.
11. Laine, B., Chartier, F., Imbert, M. and Sautiere, P. (1990) In Belaich, J.P..
Bruschi, M. and Garcia, J.L. (eds.) Microbiology and Biochemistry of strict
anaerobes involved in interspecies hydrogen transfer. Plenum Press, N.York,
pp 291-301.
12. Weintraub, H. (1973) Cold Spring Harbor Symposium Quant. Biol., 38,
247-256.
13. Lin, S.Y. and Riggs, A.D. (1974)/>roc. Noll. Acad. Sci. USA, 71, 947-951.
14. Barbier, B., Charlier, M. and Maurizot, J-C. (1984) Biochemistry, 23.
2933-2939.
15. Wolfes, H., Fliess, A., Winkler. F. and Pingoud, A. (1986) Eur. 7. Biochem.,
159, 267-273.
16. Katouzian-Safadi, M., Blazy, B. and Charlier, M. (1991) Photochem.
Photobiol. 53, 611 -616.
17. Hackett, P., and Hanawalt, P. (1966) Biochim.Biophys. Ada, 123, 356-363.
18. Imbert, M., Laine, B., Helbecque, N., Mornon, J-P., Hinichart, J-P. and
Sautiere, P. (1990) Biochim. Biophys. Acta, 1038, 346-354.
19. Le Caer, J.P. and Rossrer, J. (1988) Anal.Biochem., 169, 246-252.
20. Pandey, V.H. and Modak, M.J. (1988)7. Chromalography, 444, 133-139.
21. Varghese, A.J. (1974) Photochem. Photobiol., 20, 461-464.
22. Dietz, T.M. and Koch, T.H. (1989) Photochem. Photobiol., 49, 121-129.
23. Ito, S., Saito, I. and Matsuura, T. (1980) 7. Amer. Chem. Soc., 102,
7535-7541.
24. Saito, I., Ito, S., Matsuura, T. and Helene, C. (1981) Photochem. Photobiol.,
33, 15-19.
25 Saito, I. and Matsuura, T. (1985) Ace. Chem. Res., 18, 134- 141.
26. Dietz, T.M. and Koch, T.H. (1987) Photochem. Photobiol., 46, 971 -978.
27. Dietz, T.M., von Trebra, R.J., Swanson, B.J. and Koch, T.H. (1987) 7.
Amer. Chem. Soc., 109, 1793-1797.
28. Helene, C , Brun, F. and Yaniv, M. (1971)7. Mol. Biol., 58, 349-365.
29. Brun, F., Toulme\ J-J. and Helene, C. (1975) Biochemistry, 14, 558-563.
30. Kolodrubetz, D. (1990) Nucleic Acids Research, 18, 5565.
31. Weber, I.T., McKay, D.B. and Steitz, T.A. (1982) Nucleic Acids Research,
10, 5085-5102.
32. White, S.W., Appelt, K., Wilson, K.S. and Tanaka, I. (1989) Proteins, 5,
281-288.
33. Gaboriaud, C . Bissery, V., Benchetrit, T. and Momon, J P. (1987) FEBS
Lett., 224, 149-155.