characterization of peptidoglycan stem lengths by solid-state 13c and 15n nmr
TRANSCRIPT
Vol. 137, No. 2, 1986 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS June 13, 1986 Pages 736-741
CWARACTERIZATION OF PEPTIDOGLYCAN STEM LJINGTHS BY SOLID-STATB 13 C AND I51 NMR
Jacob Schaefer+, Joel R. Garbo”‘, Gary S. Jacob+,
Theresa M. Forrest* and G. Edwin Wilson, Jr.*
+ Monsanto Company, Physical Sciences Center, St. Louis, MO 63167
*Department of Chemistry. University of Akron, Akron, OH 44325
Received April 30, 1986
SUMMARY : Lyophilixed whole cells of Aerococcus viridans (Gaffkya homari) grown on
a synthetic medium containing D-[P- 13 15 C, N]Ala, or containing both L-[l- 13 C]Lys
and D-[l’N]Ala, have been examined by double cross-polarization magic-angle
spinning 13 C and l5 N nuclear magnetic resonance. Results from the double-labeled
alanine experiment confirm the absence of metabolic scrambling of alanine by
& viridans. Results from the combined single-label experiment can be used to
count directly the number of adjacent L-Lys and D-Ala units in peptide chains of
cell-wall peptidoglycan. This count leads to the conclusion that there are no
terminal D-Ala or D-Ala-D-Ala units in uncross-linked chains of the peptidoglycan
of A. viridans. 0 1986 Academic Press, Inc.
A component common to the cell walls of both Gram-positive and Gram-negative
bacteria is peptidoglycan. The glycan part of this polymer consists of
alternating N-acylated residues of glucosamine and its 3-0-D-lactyl ether
derivative, muramic acid (1). The peptide part of the polymer consists of short
chains, of known sequence, cross-linked to one another (Figure 1). Natural
variability in the properties of cell walls of a particular bacterium might be
achieved by alterations in the degree of cross-linking, by elimination of terminal
D-Alanine units to shorten the length of uncross-linked peptide chains, and by the
formation of n-mers (Figure 1) resulting from multiple cross-linking (1). We have
developed previously a cross-polarization magic-angle spinning (CPMAS) 15 N nuclear
magnetic resonance (NMR) technique (2) which measures directly cross-linking in
intact peptidoglycan (3). The analytical method depends upon the ability to
identify and quantify Lysyl-amino and lysyl-amide groups in intact cell walls of
0006-291X/86 $1.50 Copyright 0 1986 L~J Academic Press. Inc. All r&h& of reproduction in any form reserved. 736
Vol. 137, No. 2, 1986 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
I a I b
Figure 1. Schematic representation of the peptidoglycan of the ceLL wall of A. viridans. -- The glycan is represented by short horizontal solid lines and
the peptide units attached to the glycan by the boxes. Cross-Iinking occurs
through the r-nitrogen of lysine of one chain and the carbonyl carbon of
D-alanine of a second chain, Cross-linking between peptide chains can create
(a) dimers. (b) trimers. and (c-e) n-mers.
whole cells grown in a medium containing [t- 15 N]Lys. In this paper, we report the
characterization of the peptide chain length of intact peptidoglycan of
A. viridans (also referred to as Gaffkya homari) by 13 15 -- C- N chemical-bond labeling
(5,6) with detection by double cross-polarization (DCP) MAS 13 C NMR.
MATERIALS AND METHODS: Chemicals. D-[2- 13C 15 , Nlalanine. L-[1-13C]lysine,
D-[1-13C]alanine, and D-[15N]alanine were obtained from Merck Stable Isotopes,
Canada. All labels were 99 atom X enriched.
Culture Methods. A. viridans (ATCC 10400) was grown in the presence of labeled
amino acids in 1-2-Liter volumes (7.8). Growth with aeration and stirring was
followed optically at 660 nm. and cells were harvested at an absorbance of 0.6 by
centrifugation at 10.000 X g for 10 min. washed once with 0.025 M potassium
phosphate, pH 7.0, centrifuged again, frozen in liquid nitrogen, and lyophilized.
Magic-Angle Spinning NMR. l5 N NMR spectra were obtained at 20.3 MHz and 13c NMR
spectra at 50.2 MHz using matched spin-lock cross-polarization transfers with 2-ms
contacts and H1(C or N) = 35 kHz (9). The dried samples were contained in a
cylindrical double-bearing rotor spinning at 3.2 kHz. Residual spinning sidebands
in the spectra were suppressed by pulse techniques (10). Technical details of the
spinning and cross-polarization procedures are reported elsewhere (11.12).
Double cross-polarization 15 N NMR spectra were obtained using matched spin-lock
transfers first from 'EL to 15 N and then from 15 N to 13C (5.6). I f the 13C-rf
737
Vol. 137, No. 2. 1986 BfOCHEMlCAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
field is off resonance, the 15N signal level is SO. If the 13C-rf field is
on resonance and its amplitude satisfies the carbon-nitrogen Hartmann-Hahn
condition (6.13). a spin-lock transfer from 15 N to l3 C drains polarization
from l5 N reducing SC. The direct difference between single and double
cross-polarization experiments then results in the accumulation of a DCPMAS
difference 15 N-signal (AS) arising exclusively from those nitrogens directly
bonded to 13C . The corresponding l3 C DCPMAS spectra were obtained by first
transferring polarization from 1 H to 13 C. and then from 13 C to 15N.
RESULTS AND DISCUSSION: NMR Strategy. The strategy for the determination of -
peptide chain length in peptidoglycan is to count the number of connections
between L-Lys and D-Ala. These occur only in cell walls. The counting is
accomplished by growing bacteria in media containing L-[l- 13 C]Lys and D-[15N]Ala
and then examining intact lyophilized peptidoglycan. DCPMAS l3 C NMRcan be used
to measure quantitatively the number of 13C-15N linkages (6). The DCPMAS 13C NMR
experiment has about a six-fold sensitivity advantage over the corresponding 15N
NMR experiment. A factor of 3 of this advantage results from the higher Larmor
frequency, and a factor of 2 from the narrower linewidths usually observed. The
number of 13C-15 N linkages (together with the amino-acid composition of the cell
wall) will yield the chain length directly if there is no scrambling of L-Lys, and
if the specific isotopic enrichment of D-Ala is known. Scrambling of lysine is
not common in bacteria under our growth conditions and its absence can be
confirmed in a number of ways (3).
The specific isotopic enrichment of D-Ala can be diminished by two factors:
scrambling and de novo synthesis. -- Scrambling puts 15N label into the general
nitrogen pool where it can be routed to proteins. De novo synthesis of unlabeled --
D-Ala reduces the incorporation of labeled D-Ala into L-Lys-D-Ala linkages.
Alanine Scrambling. We measure alanine scrambling by DCPMAS 13 c NMR of
lyophilixed. whole cells of bacteria grown on media containing D-12- 13C,15N]Ala.
For A. viridans. the 13 C-labeled alanyl a-amino carbon (55 ppm) (14) has a strong
DCP difference signal (Figure 2). The DCP transfer rate for this type of 13C-15N
bond is (14 msec) -' (6). Once the natural-abundance peak intensity at 55 ppm is
subtracted. we find AS/So is 0.22. which translates to an isotopic enrichment of
738
Vol. 137, No. 2. 1986 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
3-msec hold
1"' II', "I""T"" 300 200 100 0 -100 PPM
Figure 2. 50-MHz l3 C NMR spectra of intact lyophilised cells of A. viridans --
grown in media containing D-12- 13C 15
, Nlalanine. ‘lhe double-cross polarization
difference signal shown at the top of the figure arises only from those
carbons directly bonded to 15 N, while the spectrum at the bottom of the figure
is the normal CPMAS 13 C spectrum after a 3-msec carbon spin lock.
15N of l.l+O.l (6). Thus, each 13C label has an "N neighbor and there is no
scrambling of alanine.
Alanine Synthesis. We measure the de novo synthesis of D-Ala by DCPMAS 15 N NMRof --
lyophilized cell-wall preparations from A. viridans grown on media containing
X.+-l5 N]Lys and D-[1-13C]Ala. The numb: of 15N-13C peptidyl cross-links
expected can be calculated based on the extent of alanine scrambling if any (see
above). The difference between the calculated and observed number of these
cross-links then measures the amount of de novo synthesis. For A. viridans this --
procedure led to the determination that de novo synthesis accounted for 80% of the --
D-Ala units in the cell wall (8).
Peptide Chain Length. With the levels of scrambling and de novo synthesis now in --
hand, we are in a position to determine the peptide chain length by DCPMAS 13C NMR
of whole cells grown on media containing L-[1-13C]Lys and D-[15N]Ala. For
&. viridans we measure a AS/SO of 0.026 + 0.003 (Figure 3, right). The DCPMAS 15N
difference signal is less than the noise and not detectable (Figure 3. left). The
13 C NMR AS/So determination includes a subtraction of the natural-abundance
carbonyl peak at 180 ppm, based on the spectrum of Figure 2.
We determined before that 50% of labeled L-Lys residues within A. viridans
appear in the cell wall and half of these are involved in peptidoglycan
739
Vol. 137, No. 2, 1986 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
* DCP difference ..._...n...._. x5
3-msec hold
/4
I', "1""I""I""I" 1""l""l""l"" 300 200 100 0 -100 PPII 300 200 100 0 -100 PPM
Figure 3. 20.2-M& 15N RMR (left) and 50.3-MHZ 13C NMR (right) spectra of
intact lyophilized cells of 4. viridans grown in media containing both
D-[L’N]alanine and L-[1-L3C]lysine. DCPMAS RMR spectra are shown at the top,
and CPMAS NMR spectra after a 3-msec spin lock at the bottom of the figure.
cross-links (3,151. Based on the DCP transfer rate of (5 msec) -1 for 13&5R
peptide bonds (6). we expect a AS/SO somewhere between 0.025 and 0.050, depending
on the number of additional L-Lys-D-Ala linkages within stems that are present,
over and above the 50% involved at cross-link sites. The smaller value is
calculated using 0.5x0.5x0.2x0.5, where the first factor in this product is the
expected AS/So for a L-(1- 13 C]Lys-D-[15N]Ala double-labeled bond, and the second
factor results from the assumption that there are no L-Lys-D-Ala linkages other
than those at cross-links (Figure la). The final two factors account for the 20%
incorporation of 15 N label into D-Ala (8). and for the fact that only half of the
labeled lysine within A. viridans is present in the cell wall. Labeled peptide -
bonds of L-Lys-L-Ala units in cytoplasmic protein can be safely ignored because
the low l5 N enrichment of cytoplasmic L-Ala (from racemized D-Ala. see below)
ensures a negligible DCP signal. The experimental value of AS/So of 0.026
therefore is consistent with the formation of the minimum possible number of
I-Lys-D-Ala bonds. This means there are no terminal D-Ala or D-Ala-D-Ala units in
uncross-linked chains of the peptidoglycan of &. viridans.
Based on these results and the cell-wall composition of about 1.5:l:l for
Ala:Gly:Lys (a), the average peptidoglycan peptide chain length is 3.5 (Figure 1).
The determination of whether A. viridans has a uniform distribution of dimers
(Figure la), or some fraction of an equimolar mixture of monomers and trimers
740
Vol. 137. No. 2, 1986 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
(Figures lb+lc). requires NMR techniques sensitive to the connectivity of many
spins (rather than just two), and is beyond the scope of this note. Either
possibility is consistent with the experimental amino-acid cell-wall compositional
ratios.
Alanine Racemase. Since the 15 N NMR spectrum of Figure 3 (left) shows no amine
nitrogens, there can be no significant concentration of free D-Ala-D-Ala units in
the intact cell of A. viridans. - In addition, we observe a carbonyl-carbon AS/SO
of 0.004 for A. viridans grown in the presence of D-[1-13C]Ala and D-[l'N]Ala.
This value is only 20% of that expected, if the observed levels of 13 C and 15N
alanyl residues in the cytoplasm (8) were in the form of D-Ala-D-Ala cell-wall
precursors. These levels must be due, therefore, to conversion of D-Ala to L-Ala
by a cytoplasmic racemase followed by incorporation as L-alanyl residues.
ACKNOWLEDGEMENT: This work was supported in part by Grant Number PCM-8416375 from
the National Science Foundation.
REFERENCES:
1.
2.
3.
4.
5.
6.
7.
8.
9.
10. 11.
12.
13. 14.
15.
Rogers, H.H.. Perkins, HR.. and Ward, J.B. (1980) Microbial Cell Walls and Membranes, Chapman and Hall. London, chapt. 6.
Schaefer, J. and Stejskal, E.O. (1976) J. Am. Chem. Sot. 98. 1031-1032.
Jacob, G.S., Schaefer, J.. and Wilson, G.E., Jr. (1984) J. Biol. Chem. 258, 10824-10826.
Fuchs-Cleveland, E. and Givarg. C. (19761 Proc. Natl. Acad. Sci. USA 73 4200-4204.
-'
Schaefer, J., Skokut, T.A., Stejskal, E.O., McKay, R.A., and Varner. J.E. (1981) Proc. Natl. Acad. Sci USA ;18. 5978-5982.
Schaefer, J., Stejskal. E.O.. Garbow. J.R., and McKay. R.A. (1984) J. Mag. Reson. 59, 150-156.
Miller, T.L., and Evans, J.B., (1970) J. Gen. Microbial. 61. 131-135.
Jacob, G.S.. Schaefer, J. and Wilson, G.E., Jr., (1985) J. Biol. Chem. 260, 2777-2781.
Schaefer, J. and Stejskal, E.O. (1979) High-Resolution 13 C NMR of Solid Polymers in Levy, G.C.. ed., "Topics in Carbon-13 NMR Spectroscopy," ~01.3.
Dixon, W.T. (1982) J. Chem. Phys. 77, 1800-1809. - Groombridge, C.J.. Harris, R-K.. Packer, K-J.. Say, B.J., and Tanner, S.F. (1980) J. Chem. Sot.. Chem. Comm.. 174-175. Hexem. J.G., Frey, M.H.. and Opella, S.J. (1980) J. Am. Chem. Sot. 103 -' 224-226. Hartmann, S.R., Hahn, E. L. (1962) Phys. Rev. 128, 2042-2053. Stothers. J.B. (1972) "Carbon-13 NMR Spectroecopy," Academic Press, New York, p. 479. Wilson. G.E., Jr., G.S. Jacob, and J. Schaefer (1985) Biochem. Biophys. Res. Comm. 126, 1006-1012.
741