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