AcmD, a Homolog of the Major Autolysin AcmA of Lactococcus lactis, Binds to the Cell Wall and Contributes to Cell Separation and Autolysis

Lactococcus lactis expresses the homologous glucosaminidases AcmB, AcmC, AcmA and AcmD. The latter two have three C-terminal LysM repeats for peptidoglycan binding. AcmD has much shorter intervening sequences separating the LysM repeats and a lower iso-electric point (4.3) than AcmA (10.3). Under standard laboratory conditions AcmD was mainly secreted into the culture supernatant. An L. lactis acmAacmD double mutant formed longer chains than the acmA single mutant, indicating that AcmD contributes to cell separation. This phenotype could be complemented by plasmid-encoded expression of AcmD in the double mutant. No clear difference in cellular lysis and protein secretion was observed between both mutants. Nevertheless, overexpression of AcmD resulted in increased autolysis when AcmA was present (as in the wild type strain) or when AcmA was added to the culture medium of an AcmA-minus strain. Possibly, AcmD is mainly active within the cell wall, at places where proper conditions are present for its binding and catalytic activity. Various fusion proteins carrying either the three LysM repeats of AcmA or AcmD were used to study and compare their cell wall binding characteristics. Whereas binding of the LysM domain of AcmA took place at pHs ranging from 4 to 8, LysM domain of AcmD seems to bind strongest at pH 4.


Introduction
Peptidoglycan, the major cell wall material of Gram-positive bacteria, is composed of chains of N-acetylmuramic acid and N-acetylglucosamine linked by means of β(1-4) glycosidic bonds. These bonds are cleaved by peptidoglycan hydrolases (PGH) during cell separation and cellular autolysis [8,16,25]. Impairment of cell separation activity of PGHs leads to long chains of cells in the lactic acid bacteria Lactococcus lactis and Streptococcus thermophilus [25,37,40]. Most PGHs are composed of at least two distinct domains, a cell wall-binding domain and a catalytic domain [25]. The cell wall-binding domains assist in adhering the enzymes to the murein layer while the catalytic domains cleave the cell wall. A tight interplay between both domains is essential for optimal PGH activity [42]. The (auto) lysis of the cells of lactic acid bacteria due to the action of its PGHs has been shown to be essential during cheese ripening for the release of intracellular proteins such as peptidases that contribute to flavor development [43].

General DNA techniques and transformation
Molecular cloning techniques were performed essentially as described by Sambrook et al. (23). Genomic DNA of L. lactis was isolated according to the method of Leenhouts et al. [29]. Minipreparations of plasmid DNA from L. lactis were obtained by the alkaline lysis method as described by Seegers et al. [37]. Plasmid DNA was isolated at a large scale using a Nucleobond kit PC 100 (Machery-Nagel, Düren, Germany) as specified by the supplier. Restriction enzymes, T4 DNA ligase, and deoxynucleotides were obtained from Roche Diagnostics GmbH and were used according to the supplier's instructions. Polymerase chain reactions (PCR) were performed in a Master cycler gradient (Merck KGaA) using Taq DNA polymerase or Expand DNA polymerase according to the instructions of the manufacturer (Roche Diagnostics GmbH). PCR products were purified using the High pure PCR product purification kit and protocol (Roche Diagnostics GmbH). E. coli and L. lactis were transformed by electroporation using a Gene pulser (Bio-Rad Laboratories, Richmond, CA) as described by Zabarovsky and Winburg [51] and Leenhouts and Venema [26], respectively. E. coli MC1061 cells were transformed with the recombinant vector by the heat-shock method [45].

Construction of AcmA and AcmD fusion proteins
The modular architecture of all the constructs used in this study is depicted in Figure 1. Plasmid pNGacmD [41] was digested with the restriction enzyme BamHI for which a recognition site is located in the active site of AcmD. The oligonucleotides pACMDmyc1 (GATCTAGAACAAAAACTTATTTCAGAAGAAGATCTT, underlined XbaI) and pACMDmyc2  [5] pBAD Amp r , carrying arabinose-inducible promoter P araBAD [15] pBADcLIC Amp r , pBAD derivative carrying arabinose promoter and His 10 tag [14] pBADcLIC-GFP Amp r , pBAD derivative carrying arabinose promoter and gfp fused to the His 10 tag encoding sequence [14] pBADcLIC-GFP-LysM AcmD repeats of AcmA in SwaI site [49] (GATCAAGATCTTCTTCTGAAATAAGTTTTTGTTCT, underlined XbaI) were annealed at 70 o C and cloned into the BamHI site of pNGacmD resulting in the plasmid pNGacmD::myc. Primers pACMD2 and pACMD3 (CGGAATTCAAGGAGGAGAAATATCAGGAGG; underlined EcoRI site) were used to amplify a 603-bp fragment encoding the C-terminal domain of AcmD that contains the three LysM repeats, using chromosomal DNA of L. lactis IL1403 as a template. Upon digestion with EcoRI and HindIII the fragment was cloned into the EcoRI/HindIII sites of pNG3041 thereby replacing the 1162-bp domain encoding the LysM repeats of AcmA. The resulting plasmid was named pNG3042 [5].

Construction of an acmD insertion mutant of L. lactis IL1403
pNGacmD::myc was cut with SphI and PstI. The fragment carrying acmD::myc was inserted into the SphI /PstI sites of pORI280, an integration vector which lacks the gene encoding the replication initiation protein, repA [27,28]. The resulting plasmid, pORIacmD::myc, was obtained in L. lactis strain LL108, which carries multiple copies of the repA gene on the chromosome [27]. After transformation of L. lactis IL1403acmA::ISS1 with pORIacmD::myc, erythromycin resistant integrants were checked by PCR with the primers: pEM280 (GCCCATATTTTTTCCTCC; annealing in the 5'-end of the erythromycin resistance gene) and pACMD2 (CGCAAGCTTCTGCAGAGCTCTTAGATTCTAATTGTTTGTC CTGG; underlined HindIII, which anneals to the 3'-end of acmD) as was described before [7].
Selection of the second crossover event was done as described by Leenhouts and Venema [26]. A 1040-bp region of acmD was amplified from the chromosomal DNA of selected integrants using the primers pACMD1 (CCTGTCATGAAACAGAAACATAAAT) and pACMD2, which anneal at either end of acmD. The presence of the c-myc epitope in acmD was confirmed by restriction with XbaI, as this site is only present in the DNA coding for this epitope. Although attempts were made to use the same approach to construct a single mutant of acmD in L. lactis IL1403, for unknown reasons this was not successful. Only first step integrants were obtained but excision resulted in all cases in reversion to the wild type.
LysM AcmD -GFP-His 10 protein samples were subjected to SDS-PAGE with a 15% PAA gel and in-gel GFP fluorescence was visualized using a Gel Documentation System (Bio-Rad Laboratories Inc.). Bands obtained by Western hybridization were semi-quantified using the Quantity One program (Bio-Rad Laboratories B.V., Veenendaal, the Netherlands).

Protein expression, isolation and purification
E. coli MC1061 bearing the desired plasmids (pBADcLIC-GFP-LysM AcmA or pBADcLIC-GFP-LysM AcmD ) were grown until an OD 600 of 0.6-0.8 and induced with 0.2% arabinose (Merck KGaA) for 2 h. The cells were harvested at 5000 X g for 20 min at 4° C and the cell pellet was resuspended in lysis buffer (50 mM NaH 2 PO 4 pH 8.0, 300 mM NaCl, 20 mM imidazole). Cell extract was obtained by sonication at 4° C using three cycles of 30 s pulses at 70% amplitude with 1 min intervals (Vibra Cell, Sonics & Materials, Newton, CT) followed by centrifugation at 20800 X g for 30 min at 4° C. LysM fusion protein was isolated and purified by Ni-NTA affinity chromatography as recommended by the manufacturer (Qiagen GmbH, Hilden, Germany).

Two-dimensional (2D)-gel electrophoresis
Overnight cultures of L. lactis IL1403acmA::ISS1 and IL1403acmA::ISS1acmD::myc were diluted to an OD 600 of 0.05 in fresh GM17 medium and incubated at 30° C until an OD 600 of 1.0 was reached. Subsequently, the cultures were centrifuged and proteins in the supernatant fractions were precipitated overnight at 4° C with TCA at a final concentration of 10%. Proteins were collected by centrifugation at 20000 X g for 20 min; the protein pellet was washed three times with acetone, air-dried and dissolved in a solution containing 8 M Urea, 2% CHAPS, 2 mM TBP, 0.2% ampholytes. The amount of protein loaded for 2D-gel electrophoresis was equal to that in 100 ml of supernatant fraction of GM17 cultures with an OD 600 of 1.0. Isoelectric focusing (IEF) strips (Bio-Rad Ready Strips IPG; pH 4 to 7, 11 cm) were passively rehydrated overnight with the sample to be analyzed. IEF was performed with a Bio-Rad Protean IEF cell (Bio-Rad). The voltage was step-wise increased from 150 V (0.5 h) to 300 V (1 h) and 600 V (1 h), while the voltage was raised linearly to 8000 V until 25000 Vh. Subsequently, the strips were equilibrated twice for 15 min with 5 ml equilibration buffer (0.05 M Tris-HCl pH 8.8, 6 M Urea, 30% (w/v) glycerol, 2% SDS), the first time containing with 1% DTT, the second time with 4% iodoacetamide. The equilibrated strips were loaded on a standard SDS-(12%) PAA gel and run for 2.5 h at 100 V and 15° C. The proteins in the gel were fixed by washing with a solution of 40% ethanol and 10% acetic acid. Staining was performed overnight with colloidal Coomassie brilliant blue (0.1% CBB G-250, 2% phosphoric acid, 20% ethanol); destaining was with distilled water. PDQuest 2D analysis software (Bio-Rad Laboratories Inc.) was used for comparison between gels and analysis of the 2D gel data.

Bacterial growth, enzyme assays, and microscopy
OD 600 's of cultures were measured in a Novaspec II spectrophotometer (Pharmacia Biotech AB, Uppsala, Sweden). To measure cellular lysis of L. lactis, cells were grown in GM17 at 30° C for 50 h. Subsequently, as a measure of the extent of culture lysis, release of intracellular X-prolyl dipeptidyl aminopeptidase (PepX) was measured using the chromogenic substrate Ala-Pro-p-nitroanilid (Bachem Feinchemicalien AG, Bubendorf, Switzerland) as described earlier [42].
The sample/substrate mixture was pipetted into a microtiter plate well, and color development was monitored in a THERMOmax microtiter plate reader (Molecular Devices Corporation, Menlo Oaks, CA) at 405 nm for 145 min at 37° C. The slope of the substrate hydrolysis/color development was calculated for two independent experiments.
Light microscopy pictures of L. lactis were made with a Zeiss microscope (Carl Zeiss, Thornwood, CA) and an Axiovision digital camera (Axion Technologies, Houston, TX). A fluorescence microscope (Zeiss Axiophot) fitted with a digital camera and a green filter was used to view GFP fluorescence.
For electron microscopic analysis, L. lactis MG1363 cells were treated with TCA (as described below) and incubated with MSA2, MSA2-LysM AcmA or MSA2-LysM AcmD as described above for untreated cells. The antibodies against MSA2 were diluted 1:1000 in PBS-containing 0.15 M glycine. Immunogold labeling was performed using Auroprobe 15 nm goat anti-rabbit IgG gold marker (Amersham Biosciences) using preparations of glutaraldehyde-fixed cells on Formvar-carbon-coated nickel grids. The labeled samples were stained with 0.1% uranyl acetate (w/v in water) and examined in a Philips CM10 transmission electron microscope at 100 kV.
To measure the influence of AcmD on cellular lysis, cells of L. lactis NZ9000 (pNGacmD) or NZ9000acmAΔ1 (pNGacmD), both with or without nisin-induced AcmD, were mixed with AcmA and/or AcmD. Cells from 10 ml of culture of L. lactis NZ9000 (pNGacmD) or NZ9000acmAΔ1 (pNGacmD), both either or not induced with nisin at an OD 600 of 0.6, were collected by centrifugation at 5000 X g 2 h after induction. The cell pellets were resuspended in 10 ml of the supernatants of the L. lactis MG1363 (which contains both AcmA and AcmD) or L. lactis MG1363acmAΔ1 (which contains AcmD) cultures and incubated at 30° C for 2.5 h (OD 600 ~1.8). Subsequently, supernatant samples were collected and used to measure the extent of culture lysis by measuring PepX activity.

Binding of fusion proteins to lactococcal cells
MSA2-LysM AcmA and MSA2-LysM AcmD binding studies were performed by mixing equal amounts of L. lactis cells (the amount of cells present in 1 ml of culture with an OD 600 of 1.0) or with L. lactis cells that had been treated with trichloroacetic acid by boiling 25 µl of cell culture in 1 ml of 10% trichloroacetic acid (TCA) with 1 ml of supernatant of a nisin-induced L. lactis NZ9000 acmAΔ1 (pNG3041, producing MSA2-LysM AcmA ) or NZ9000 acmAΔ1 (pNG3042, expressing MSA2-LysM AcmD ) cultures. The suspensions were incubated at room temperature for 5 min, centrifuged (1 min at 20000 X g) and washed once with M17 broth. The pellets were subsequently resuspended in SDS sample buffer, boiled for 5 min, and subjected to SDS-PAGE.
Purified LysM AcmA -GFP-His 10 and LysM AcmD -GFP-His 10 purified proteins (3 µM each) were added to L. lactis cells in 50 mM NaH 2 PO 4 (pH 4.0, 6.0 and 8.0), 50 mM NaCl buffer at room temperature and incubated for 30 min. Cells were centrifuged at 20000 X g for 1 min and the pellet was washed three times with the same buffer containing 150 mM NaCl to remove un-specifically bound proteins. Finally the washed pellet was resuspended in the same buffer and observed under a fluorescence microscope. To rule out the possibility of GFP interaction with lactis cells, GFP, purified by Hydrophobic Interaction Chromatography was used as a control.

Comparison of AcmA and AcmD
A comparison between AcmA [GenBank AAK04370] and AcmD [GenBank AAK04639] of L. lactis IL1403 was performed to pinpoint the differences between both proteins. Both have a similar modular structure ( Figure 1): a signal sequence is followed by an active site domain (Glucosaminidase family, pfam PF01832) and a C-terminal LysM domain with three LysM repeats (pfam PF01476). The homology between the two proteins is around 58%. AcmD differs from AcmA in the length of the signal sequence (26 versus 57 amino acid residues), the presence of shorter amino acid sequences separating the repeats, while the protein has a pI of 4.3 instead of 10.3 for AcmA ( Figure 1 and Table 1). His-tagged purified AcmD showed PGH activity only at pH 4 while AcmA is active at pH 4 to 8 [16,43].
AcmD of L. lactis subsp. lactis strains KF147 [38] and IL1403 [4] are identical while those of strains IO-1 [19] and CV56 [12] are 99% identical. The AcmD proteins of the L. lactis subsp. cremoris A76 [3], UC509.9 [1], MG1363 [50] and SK11 [32] are around 95% identical to IL1403 AcmD, respectively. The differences are mainly located between the active site domain and the LysM domain and in the intervening sequences separating the LysM repeats. No proteins homologous to AcmD that share the same low pI are encoded by the genomes of the other bacteria sequenced to date.

AcmD contributes to cell separation and autolysis
A c-myc epitope was inserted in frame in the active site of AcmD to be able to follow expression and localization of the AcmD fusion protein and to make an acmD mutant by chromosomal integration. The chromosomal acmD gene in IL1403acmA::ISS1 was replaced by acmD::myc via replacement recombination, resulting in an acmAacmD double mutant. However for yet unknown reasons the integration of acmD::myc into the genome of IL1403 failed. The effects of mutation of acmD on protein secretion by and autolysis and cell separation of L. lactis were investigated. Activity of AcmD could not be detected when cell-free extracts of L. lactis NZ9000, L. lactis IL1403, or IL1403acmA::ISS1 all carrying an intact copy of the acmD gene in their chromosomes, were subjected to zymography using cell wall fragments of M. lysodeikticus or L. lactis, even after renaturation of the proteins in the gel at pH 4 (results not shown).
The protein patterns of culture supernatant fractions from the acmA and acmAacmD mutants did not differ significantly, as judged by 2D-gel electrophoresis. Two proteins were present in higher amounts in either the supernatant of the acmA strain or the acmAacmD double mutant while 4 proteins were detected only in the supernatant of the acmA mutant. The main difference was the expected shift in size of the secreted AcmD protein in the double mutant as a consequence of the c-myc epitope (from 34.9 to 36.6 kDa; Figure S1). Cellular autolysis was examined by following the release of the cytoplasmic peptidase Pep-X into the culture supernatant [9]. No difference in lysis was observed between the acmAacmD double mutant and the acmA single mutant of L. lactis IL1403 ( Figure S2).
Nisin-induced overexpression of AcmD has previously been shown to result in increased lysis of L. lactis MG1363 in the presence of AcmA while no additional lysis was obtained for the empty plasmid control [43]. Knowing that AcmD is secreted into the culture supernatant even though no activity could be detected, we investigated whether the contribution of AcmD activity to lysis occurs by peptidoglycan degradation during passage of the protein through the cell wall. Because a single acmD mutant could not be obtained the following approach was used to examine AcmD-mediated lysis: L. lactis strains NZ9000 (pNGacmD) and NZ9000acmAΔ1 (pNGacmD) were grown in the presence of nisin to induce the expression of AcmD. Cell pellets were collected two hours after induction and resuspended in cell free supernatants of L. lactis MG1363, containing both AcmA and AcmD, or L. lactis MG1363acmAΔ1, containing only AcmD. Un-induced cells were used as controls. Using this approach the influence of AcmD when produced from the inside or added from the outside could be separately measured. The cells of the strains NZ9000 (pNGacmD) and NZ9000acmAΔ1 (pNGacmD) released significantly more PepX and thus lysed to a larger extent when AcmD expression was induced (Figure 2, compare un-induced with induced). This was only the case when AcmA activity was present, upon suspension of the cells in the AcmA-containing MG1363 supernatant (Figure 2, see dark grey bars), showing that AcmD contributes to autolysis by its action within the cell wall.
After overnight growth at 30° C in GM17 broth differences in cell sedimentation were observed between the L. lactis strains IL1403, IL1403acmA::ISS1 and IL1403acmA::ISS1acmD::myc. The wild type IL1403 strain did not sediment whereas the acmA mutant did and the acmAacmD double mutant even more ( Figure 3). Light microscopic analysis revealed that the latter strain formed very long chains in comparison to the other two ( Figure 3B). The chains formed by the double mutant were calculated to be on average 2 to 3 times longer than those of IL1403acmA::ISS1 ( Figure 3D). Further, complementation of AcmD expression in the acmAacmD double mutant resulted in a decrease in the length of the chains to a length resembling that of the acmA single mutant ( Figure 3C and D). This data indicates that AcmD, like AcmA, is involved in cell separation.

LysM AcmD binds to lactococcal cells preferably at low pH
As peptidoglycan binding of AcmD is difficult to examine because its activity is not detectable and antibodies are not available, LysM AcmD (pI=4.2, see column pI LysM in Table 1 and Figure 1) was fused to the MSA2 reporter protein, for which an antibody is available [35]. The fusion protein, MSA2-LysM AcmD , is secreted by fusing the signal sequence of the lactococcal proteinase PrtP to its N-terminal. Western detection revealed that MSA2-LysM AcmD is secreted from L. lactis NZ9000acmAΔ1 (pNG3042) upon nisin induction (results not shown). When a supernatant containing MSA2-LysM AcmD was mixed with L. lactis MG1363 cells poor binding of the fusion protein was observed at pH 6.2 ( Figure 4A). A similar result was obtained when L. lactis MG1363 cells were mixed with the supernatant of L. lactis NZ9000acmAΔ1 (pNG3041), containing the MSA2-LysM AcmA fusion protein [41]. Although previously we have shown that MSA2 alone cannot bind lactococcal cells [41] we did observe binding of the MSA2 protein in this study. When the binding of the three proteins was repeated with TCApretreated L. lactis cells a substantial increase in binding was observed for the MSA2-LysM AcmA and MSA2-LysM AcmD fusion proteins but not for the MSA2 protein ( Figure 4A). To investigate whether the pI of MSA2-LysM AcmD affects cell binding the binding to TCA-pretreated cells was repeated at pHs 3.2 and 6.2. Figure 4B clearly shows that MSA2-LysM AcmD binds 2-fold more at pH 3.2 than at pH 6.2 (compare lanes 1 and 3). No detectable binding of MSA2-LysM AcmA could be observed using electron microscopy ( Figure 4C) while strong labeling signals was detected on TCA-treated cells. However, as for MSA2-LysM AcmD , only minor labeling signals were observed ( Figure 4C). These results suggest that LysM AcmD has poor cell wall binding properties around neutral pH (pH 7.2) while its binding to the cell wall is increased at a pH closer to its pI, when the net charge of the protein is positive.

LysM domains of AcmD and AcmA bind differently to lactococcal cells
The purified MSA2-LysM AcmA fusion protein binds to the septum and poles of L. lactis cells when added from the outside to these cells [41]. Attempts to repeat these experiments with the MSA2-LysM AcmD protein failed due to side effects of the immunodetection at pH 4. And because the MSA2 protein showed non-specific binding to L. lactis cells another approach was taken. To locate the sites of binding of LysM AcmD on the cell, C-terminal GFP fusions of His 10 -tagged LysM AcmA (pI=10, see column pI LysM in Table 1) and LysM AcmD were expressed in and purified from E. coli using Ni-NTA. SDS-PAGE with a 15% PAA gel followed by in-gel fluorescence detection showed that the fusion proteins are of the proper size and fluoresce (results not shown and Figure S3). Purified LysM AcmD -GFP-His 10 (48.5 kDa; pI=5.6) only binds to L. lactis NZ9000 cells at pH 4, not at pH 6 or pH 8 ( Figure 5A and Figure S5). Phase contrast and fluorescence analyses showed that L. lactis does not intrinsically fluorescence under the conditions used ( Figure S4A). HIC-purified GFP protein did not bind to the L. lactis cells under any of the conditions tested ( Figure S4B and results not shown). Unlike the LysM AcmD fusion protein, LysM AcmA -GFP-His 10 (51.4 kDa; pI=8.2) bound to L. lactis cells at pH 4, pH 6 and pH 8 ( Figure 5B and C and results not shown).

Discussion
In this study we show that the PGH AcmD of L. lactis is involved in cell separation as deletion of acmD results in increased length of the chains of cells. AcmD also contributes to autolysis, but only in the presence of AcmA. The LysM domain of AcmD binds more to the lactococcal cell wall at a pH close to its pI.
Under standard laboratory growth conditions at pH 6 to 7, AcmD is produced and mainly secreted by L. lactis ( Figure S1). Peptidoglycan degrading activity could not be detected for the native expressed or overexpressed AcmD enzyme using a zymographic analysis method (results not shown). Earlier, Histagged purified AcmD protein was shown to be active under renaturation conditions also employed in the present study (pH 4) [16]. A possible explanation for the different results in the two studies could be the very high protein concentration that was used in the earlier study. Although we were not able to show in vitro AcmD activity, a clear phenotypic effect was seen in vivo. The L. lactis acmAacmD mutant formed significantly longer chains ( Figure 3B and D) and sedimented more readily than the wild type strain ( Figure 3A) showing that AcmD is involved in cell separation, like its homolog AcmA. Complementation of the acmD mutation by plasmid-specified AcmD resulted in a decrease of chain length of the acmAacmD mutant ( Figure 3C).
A second phenotypic effect was obtained upon overexpression of AcmD, namely increased autolysis of L. lactis (Figure 2). This effect was only observed when AcmA activity was also present: autolysis was not increased when AcmD was overexpressed in an AcmA-minus background. Although activity and presence in the cell wall of AcmD could not be unequivocally detected, it is possible that AcmD hydrolyses the peptidoglycan at specific sites during its passage through the cell wall. It has been shown that a pHgradient is present over the cell wall of the Gram-positive model bacterium Bacillus subtilis and that this affects the activity of enzymes such as PGH's [20]. Possibly, such a pH gradient also exists in the lactococcal cell wall, which would generate a local low pH required for AcmD activity.
The LysM AcmA -GFP fusion protein seems to bind at all pH's tested to the whole surface of lactococcal cells ( Figure 5B and C). Previously, hotspots for binding of the MSA2-LysM AcmA fusion protein were detected by immunofluorescence microscopy [41]. This discrepancy may be caused by differences in the N-(MSA2) and C-terminal (GFP) fusion proteins used in the two studies e.g., the sizes of the proteins used, or by differences in the limits of detection of immunofluorescence and GFP fluorescence. Whereas the GFP fusion proteins can be detected in and (slightly) outside the cell wall, detection of the MSA2 fusion protein depends on proper outer surface exposure of the epitopes that are to be bound by the antibodies, which are too big to enter the cell wall [11]. LysM AcmD -GFP fusion protein (48.5 kDa; pI=5.6) binds to the cell wall at pH 4, when its net charge is slightly positive. At a pH above or below their pI, proteins carry either a net negative or positive charge, respectively, and local pH can greatly influence the characteristics of proteins or their domains. Although AcmA and AcmD share amino acid sequence homology in their active sites and their cell wall-binding domains, a significant difference exists in the pI values of their LysM domains. The observed differences in binding of the LysM domains of AcmA and AcmD ( Figure 5) might be due to differences in the LysM domain pI's and/or the intervening amino acid sequences in these domain. It has been suggested that the low pI of some of the LysM-containing proteins functions in binding of these enzymes to the cell wall at low pH or for positioning of active site domains at the proper location in the peptidoglycan matrix close to the membrane [8]. A more detailed study is needed to determine the specific binding sites in the cell wall peptidoglycan layer for the various types of LysM domains and whether the domains may be used to position different PGHs to different sites in the cell wall. Figure S1. Comparison of 2D-gel images of supernatant fractions of L. lactis IL1403acmA::ISS1 and L. lactis IL1403acmA::ISS1acmD::myc.

Supporting Information
xmlns:xlink="http:// www.w3.org/1999/xlink" xmlns:mml="http://www.w3.org/1998/ Math/MathML">The amount of protein loaded in both cases was the equivalent of supernatant fraction of 100 ml of a GM17 culture with an optical density at 600 nm of 1.0. The position of the spots of the AcmD and AcmD::myc proteins, identified by Mass-spectroscopic analysis, and their molecular weights and pIs are indicated. Proteins that were more abundant in the supernatant fraction of IL1403acmA::ISS1 (blue) or IL1403acmA::ISS1acmD::myc (red), and those unique in the supernatant of IL1403acmA::ISS1 (yellow) or IL1403acmA::ISS1acmD::myc (green) are indicated. Sizes of the pre-stained molecular mass marker (kDa) are indicated in the middle. (TIF) Figure S2. Deletion of acmD does not affect cell lysis during growth. Release of intracellular X-prolyl dipeptidyl aminopeptidase (PepX) from L. lactis IL1403 (♦), IL1403acmA::ISS1 (▲) and IL1403acmA::ISS1acmD::myc (■). Samples were taken at the indicated time points from the bacterial cultures incubated in GM17 broth. Upon removal of the cells by centrifugation the PepX-activity (in arbitrary units) released into the medium due to autolysis was determined using a chromogenic substrate, as described in the Materials and Methods section. (TIF) Figure S3. Expression of LysM AcmD -GFP-His 10 in E. coli. Coomassie brilliant blue-stained SDS-(15%) PAA-gel (left) and in-gel GFP-fluorescence (right) showing the expression of LysM AcmD -GFP-His 10 on SDS-PAGE with a 15% PAA gel. E. coli MC1061 bearing the pBADcLIC-LysM AcmD was grown at 37°C until OD 600 of 0.8 and induced with 0.2% arabinose for 2 h (see Materials and Methods section). The cell extracts of control and test samples were loaded on PAA gel for the identification of specific protein band. For the latter figure, the PAA gel is exposed to UV-light prior to coomassie staining for imaging the fluorescent bands. Prestained protein marker lane 1, cell extracts of empty vector control strain, un-induced control and 0.2%-arabinose induced test samples in lanes 2, 3 and 4, respectively. Arrows indicate LysM AcmD -GFP-His 10 protein/activity bands. (TIF)