Portal de Periódicos da CAPES

The Solution Structure of Acyl Carrier Protein from Mycobacterium tuberculosis

2002; Elsevier BV; Volume: 277; Issue: 18 Linguagem: Inglês

10.1074/jbc.m112300200

1083-351X

Hing C. Wong, Gaohua Liu, Yongmei Zhang, Charles O. Rock, Jie Zheng,

Bacteriophages and microbial interactions

Acyl carrier protein (ACP) performs the essential function of shuttling the intermediates between the enzymes that constitute the type II fatty acid synthase system. Mycobacterium tuberculosis is unique in producing extremely long mycolic acids, and tubercular ACP, AcpM, is also unique in possessing a longer carboxyl terminus than other ACPs. We determined the solution structure of AcpM using protein NMR spectroscopy to define the similarities and differences between AcpM and the typical structures. The amino-terminal region of the structure is well defined and consists of four helices arranged in a right-handed bundle held together by interhelical hydrophobic interactions similar to the structures of other bacterial ACPs. The unique carboxyl-terminal extension from helix IV has a “melted down” feature, and the end of the molecule is a random coil. A comparison of the apo- and holo-forms of AcpM revealed that the 4′-phosphopantetheine group oscillates between two states; in one it is bound to a hydrophobic groove on the surface of AcpM, and in another it is solvent-exposed. The similarity between AcpM and other ACPs reveals the conserved structural motif that is recognized by all type II enzymes. However, the function of the coil domain extending from helix IV to the carboxyl terminus remains enigmatic, but its structural characteristics suggest that it may interact with the very long chain intermediates in mycolic acid biosynthesis or control specific protein-protein interactions. Acyl carrier protein (ACP) performs the essential function of shuttling the intermediates between the enzymes that constitute the type II fatty acid synthase system. Mycobacterium tuberculosis is unique in producing extremely long mycolic acids, and tubercular ACP, AcpM, is also unique in possessing a longer carboxyl terminus than other ACPs. We determined the solution structure of AcpM using protein NMR spectroscopy to define the similarities and differences between AcpM and the typical structures. The amino-terminal region of the structure is well defined and consists of four helices arranged in a right-handed bundle held together by interhelical hydrophobic interactions similar to the structures of other bacterial ACPs. The unique carboxyl-terminal extension from helix IV has a “melted down” feature, and the end of the molecule is a random coil. A comparison of the apo- and holo-forms of AcpM revealed that the 4′-phosphopantetheine group oscillates between two states; in one it is bound to a hydrophobic groove on the surface of AcpM, and in another it is solvent-exposed. The similarity between AcpM and other ACPs reveals the conserved structural motif that is recognized by all type II enzymes. However, the function of the coil domain extending from helix IV to the carboxyl terminus remains enigmatic, but its structural characteristics suggest that it may interact with the very long chain intermediates in mycolic acid biosynthesis or control specific protein-protein interactions. acyl carrier protein ACP from M. tuberculosis 4′-phosphopantetheine holo-(ACP) synthase nuclear Overhauser effect nuclear Overhauser effect spectroscopy heteronuclear single quantum coherence 4-morpholinepropanesulfonic acid dithiothreitol There are two types of fatty acid synthase systems. The type I or “hard-wired” system is found in metazoans and is carried out by a multifunctional polypeptide with multiple active sites (1Smith S. FASEB J. 1994; 8: 1248-1259Crossref PubMed Scopus (508) Google Scholar). In contrast, the type II system found in bacteria and plants consists of a set of discrete monofunctional proteins, each encoded by a separate gene (2Cronan Jr., J.E. Rock C.O. Neidhardt F.C. Curtis R. Cross C.A. Ingraham J.L. Lin E.C.C. Low K.B. Magasanik B. Reznikoff W. Riley M. Schaechter M. Umbarger H.E. Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology. American Society for Microbiology, Washington, D. C.1996: 616-636Google Scholar, 3Rock C.O. Cronan J.E. Biochim. Biophys. Acta. 1996; 1302: 1-16Crossref PubMed Scopus (287) Google Scholar). ACP1 is central to both of these pathways because it functions to ferry the pathway intermediates between active site centers or enzymes. ACPs are also critical to the function of other metabolic pathways such as polyketide synthases (4Hopwood D.A. Sherman D.H. Annu. Rev. Genet. 1990; 24: 37-66Crossref PubMed Google Scholar). The type II fatty acid synthase ACPs are abundant, small, acidic proteins that carry the acyl intermediates attached as thioesters to the terminus of the 4′-PP prosthetic group (2Cronan Jr., J.E. Rock C.O. Neidhardt F.C. Curtis R. Cross C.A. Ingraham J.L. Lin E.C.C. Low K.B. Magasanik B. Reznikoff W. Riley M. Schaechter M. Umbarger H.E. Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology. American Society for Microbiology, Washington, D. C.1996: 616-636Google Scholar, 3Rock C.O. Cronan J.E. Biochim. Biophys. Acta. 1996; 1302: 1-16Crossref PubMed Scopus (287) Google Scholar). This prosthetic group is added post-translationally to apoACP by holo-(acyl carrier protein) synthase (AcpS), which transfers the 4′-PP moiety of CoA to Ser-36 of apoACP. There are now over a hundred highly similar ACP primary sequences in the molecular data bases and the NMR solution structures of Escherichia coli (Gram-negative) (5Holak T.A. Nilges M. Prestegard J.H. Gronenborn A.M. Clore G.M. Eur. J. Biochem. 1988; 175: 9-15Crossref PubMed Scopus (72) Google Scholar,6Kim Y. Prestegard J.H. Biochemistry. 1989; 28: 8792-8797Crossref PubMed Scopus (121) Google Scholar) and Bacillus subtilis (Gram-positive) (7Xu G.Y. Tam A. Lin L. Hixon J. Fritz C.C. Powers R. Structure. 2001; 9: 277-287Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar) ACPs are known, and there is a crystal structure of ACP bound to AcpS (8Parris K.D. Lin L. Tam A. Mathew R. Hixon J. Stahl M. Fritz C.C. Seehra J. Somers W.S. Structure. 2000; 8: 883-895Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar). There is also a NMR structure for the apoACP involved in the actinorhodin polyketide synthase from Streptomyces coelicolor (9Crump M.P. Crosby J. Dempsey C.E. Parkinson J.A. Murray M. Hopwood D.A. Simpson T.J. Biochemistry. 1997; 36: 6000-6008Crossref PubMed Scopus (141) Google Scholar), an ACP similar to the type II fatty acid synthase ACPs. These proteins are very similar and are composed of a four-α-helix bundle with the prosthetic group attached to a conserved Ser-36 located in helix II. This helix is proposed to be the site for interaction of ACP with target proteins based on the analysis of mutant proteins (10Zhang Y.M. Rao M.S. Heath R.J. Price A.C. Olson A.J. Rock C.O. White S.W. J. Biol. Chem. 2001; 276: 8231-8238Abstract Full Text Full Text PDF PubMed Scopus (147) Google Scholar), and indeed helix II is the portion of ACP that associates with AcpS in the crystal structure and becomes distorted to position Ser-36 to accept the incoming prosthetic group (8Parris K.D. Lin L. Tam A. Mathew R. Hixon J. Stahl M. Fritz C.C. Seehra J. Somers W.S. Structure. 2000; 8: 883-895Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar). An extended loop connects helices I and II; α helix III is short and was not detected as a structural element in the earliest NMR structure determinations (5Holak T.A. Nilges M. Prestegard J.H. Gronenborn A.M. Clore G.M. Eur. J. Biochem. 1988; 175: 9-15Crossref PubMed Scopus (72) Google Scholar, 9Crump M.P. Crosby J. Dempsey C.E. Parkinson J.A. Murray M. Hopwood D.A. Simpson T.J. Biochemistry. 1997; 36: 6000-6008Crossref PubMed Scopus (141) Google Scholar, 11Okada M. Matsuzaki H. Shibuya I. Matsumoto K. J. Bacteriol. 1994; 176: 7456-7461Crossref PubMed Google Scholar). The high degree of similarity among these structures is consistent with the ACPs from one species being used efficiently by the fatty acid synthase enzymes from another species. In Mycobacterium tuberculosis, AcpM is one of the three proteins that have an ACP signature sequence (12Cole S.T. Brosch R. Parkhill J. Garnier T. Churcher C. Harris D. Gordon S.V. Eiglmeier K. Gas S. Barry III, C.E. Tekaia F. Badcock K. Basham D. Brown D. Chillingworth T. Connor R. Davies R. Devlin K. Feltwell T. Gentles S. Hamlin N. Holroyd S. Hornsby T. Jagels K. Barrell B.G. et al.Nature. 1998; 393: 537-544Crossref PubMed Scopus (6406) Google Scholar). However, inMycobacterium leprae, which has the smallest genome among the mycobacteria (13Cole S.T. Eiglmeier K. Parkhill J. James K.D. Thomson N.R. Wheeler P.R. Honore N. Garnier T. Churcher C. Harris D. Mungall K. Basham D. Brown D. Chillingworth T. Connor R. Davies R.M. Devlin K. Duthoy S. Feltwell T. Fraser A. Hamlin N. Holroyd S. Hornsby T. Jagels K. Lacroix C. Maclean J. Moule S. Murphy L. Oliver K. Quail M.A. Rajandream M.A. Rutherford K.M. Rutter S. Seeger K. Simon S. Simmonds M. Skelton J. Squares R. Squares S. Stevens K. Taylor K. Whitehead S. Woodward J.R. Barrell B.G. Nature. 2001; 409: 1007-1011Crossref PubMed Scopus (1352) Google Scholar), AcpM is the only ACP-like protein. The function of AcpM in M. tuberculosis type II fatty acid synthase system is supported by the location of the acpM (Rv2244) in an operon with other genes that encode pathway enzymes (12Cole S.T. Brosch R. Parkhill J. Garnier T. Churcher C. Harris D. Gordon S.V. Eiglmeier K. Gas S. Barry III, C.E. Tekaia F. Badcock K. Basham D. Brown D. Chillingworth T. Connor R. Davies R. Devlin K. Feltwell T. Gentles S. Hamlin N. Holroyd S. Hornsby T. Jagels K. Barrell B.G. et al.Nature. 1998; 393: 537-544Crossref PubMed Scopus (6406) Google Scholar), the identification of mycolic acid precursors bound to AcpM (14Yuan Y. Mead D. Schroeder B.G. Zhu Y. Barry III, C.E. J. Biol. Chem. 1998; 273: 21282-21290Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar), and the cross-linking of AcpM to the elongation condensing enzyme, KasA (15Barry III, C.E. Lee R.E. Mdluli K. Sampson A.E. Schroeder B.G. Slayden R.A. Yuan Y. Prog. Lipid Res. 1998; 37: 143-179Crossref PubMed Scopus (442) Google Scholar). AcpM is highly expressed in E. coli and is isolated as a mixture of apoAcpM, AcpM, and AcpM acylated by long-chain fatty acids, primarily palmitate (data not shown) (16Schaeffer M.L. Agnihotri G. Kallender H. Brennan P.J. Lonsdale J.T. Biochim. Biophys. Acta. 2001; 1532: 67-78Crossref PubMed Scopus (36) Google Scholar, 17Kremer L. Nampoothiri K.M. Lesjean S. Dover L.G. Graham S. Betts J. Brennan P.J. Minnikin D.E. Locht C. Besra G.S. J. Biol. Chem. 2001; 276: 27967-27974Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). Although AcpM functions with the type II enzymes in E. coli (16Schaeffer M.L. Agnihotri G. Kallender H. Brennan P.J. Lonsdale J.T. Biochim. Biophys. Acta. 2001; 1532: 67-78Crossref PubMed Scopus (36) Google Scholar), it is unique compared with all other bacterial type II ACPs. The protein is similar to the other ACPs in the amino terminus but contains an extended carboxyl terminus of about 35 residues for which the structure and function are unknown. In this report, we define the solution structure of AcpM by protein NMR spectroscopy. The goals are to determine whether this novel ACP has a core structure similar to the typical bacterial ACPs, to define the structure of the unique carboxyl-terminal extension on AcpM, and to gain insight into the dynamics of the 4′-PP prosthetic group, which the published ACP NMR structures have not addressed. The acpMgene was amplified from M. tuberculosis λgt11 genomic library using primer pair 5′-CATATGCCTGTCACTCAGGAAG-3′ and 5′-GGATCCAAGGCTGACTCACTT-3′ to introduce a NdeI site at the initiator methionine and a BamHI site downstream the stop codon. The PCR product was ligated into pCR2.1 (Invitrogen). The clone with the correct DNA sequence was digested withNdeI and BamHI and ligated into the pET11a vector. The uniform 15N or15N/13C-labeled AcpM was expressed in E. coli strain BL21 CodonPlus (DE3)-RP growing in MOPS minimal medium containing 0.1% 15NH4Cl or 0.1%15NH4Cl and 0.2%d-[U-13C]glucose. Cell pellet was resuspended in buffer containing 20 mm Tris-HCl, pH 7.5, 1 mm EDTA, and 1 mm DTT. Cells were lysed by two passages through a French pressure cell at 20,000 p.s.i. The AcpM protein in the cell-free extract was purified using gradient elution from a DEAE-cellulose column followed by gel filtration on a HiLoad 26/60 Superdex 75 column. The purified protein sample was subjected to mass spectral analysis and determined to be a mixture of apoAcpM, AcpM, acyl-AcpM (acylated with a variety of saturated long-chain fatty acids consisting of 14, 16, and 18 carbon acyl chains), and AcpM dimers. To get a sample containing only apoAcpM and AcpM for NMR spectroscopy, protein purified by gel filtration chromatography was treated with 100 mm DTT at 37 °C overnight to reduce the dimer to monomer AcpM, and to remove acyl chains, converting acyl-AcpM to AcpM. A precipitate formed during the reaction because of the formation of acyl-DTT, which is insoluble. The apo- and holoforms were separated from the precipitate by centrifugation and dialyzed against 20 mm potassium phosphate buffer, pH 6.5, at 4 °C overnight. E. coliHis-tagged, purified AcpS protein was used to convert apoAcpM to holoAcpM. The reaction contained 2 mm AcpM (a mixture of apo- and holoforms), 10 mm CoA, 10 mm DTT, 10 mm MgCl2, and 3.6 μg of E. coliAcpS in 20 mm potassium phosphate buffer, pH 6.5. After incubation at 37 °C overnight, the AcpM was repurified. Two types of AcpM samples were generated for the structural studies. One was 15N-labeled, and another one was dual labeled with 15N/13C. The15N/13C-labeled sample was converted to 100% holoAcpM as described above, and the 15N-labeled sample was a mixture of apo- (about 40%) and holoforms (about 60%). Because15N/13C-labeled protein was synthesized using an 15N/13C enriched medium and unlabeled CoA was used to convert apoAcpM to AcpM, only about 40% of the bound 4′-PP groups in the sample were 15N/13C-labeled, and the rest of them were unlabeled. Protein samples were dialyzed against 40 mm potassium phosphate buffer, pH 6.5, containing 0.1 mm DTT and 0.1% NaN3 at 4 °C. The dialyzed samples were concentrated to about 2 mm for NMR spectroscopy studies. All NMR data were acquired with Varian Inova 600-MHz spectrometers at 27 °C. Data were processed and displayed by the program packages NMRpipe and NMRDraw (18Delaglio F. Grzesiek S. Vuister G.W. Zhu G. Pfeifer J. Bax A. J. Biomol. NMR. 1995; 6: 277-293Crossref PubMed Scopus (11280) Google Scholar) on an SGI Octane work station. The programs XEASY (19Xia T.-H. Bartels C. Wuthrich K. XEASY ETH Automated Spectroscopy for X Window System, User Mannal. ETH-Honggerberg, Zurich1993Google Scholar) and CSI (20Wishart D.S. Sykes B.D. J. Biomol. NMR. 1994; 4: 171-180Crossref PubMed Scopus (1893) Google Scholar) were used for data analysis and semiautomatic assignments. Backbone resonances were assigned on the basis of three-dimensional HNCA, HNCACB, CBCA(CO)NH, HNCO, and HNCOCA, and the automated program Assign2 (21Lukin J.A. Gove A.P. Talukdar S.N. Ho C. J. Biomol. NMR. 1997; 9: 151-166Crossref PubMed Scopus (83) Google Scholar) was used. Side-chain resonances were assigned on the basis of three-dimensional 15N-edited TOCSY, HCCH-COSY, and HCCH-TOCSY spectra. The NOE connections were assigned on the basis of three-dimensional 15N-edited NOESY and13C-edited NOESY with the help of four-dimensional15N/13C NOESY and four-dimensional13C/13C HMQC-NOESY-HSQC. An automated program, NOAH (22Mumenthaler C. Guntert P. Braun W. Wuthrich K. J. Biomol. NMR. 1997; 10: 351-362Crossref PubMed Scopus (134) Google Scholar), was used for the NOE assignment and all the assignments were manually checked. A total of 2043 meaningful distance constraints were derived from NMR data. Integrated NOE peaks were calibrated and converted to distance constraints with the program CALIBA (23Guntert P. Braun W. Wuthrich K. J. Mol. Biol. 1991; 217: 517-530Crossref PubMed Scopus (911) Google Scholar). In the final structural determination, the program DYANA (24Guntert P. Mumenthaler C. Wuthrich K. J. Mol. Biol. 1997; 273: 283-298Crossref PubMed Scopus (2537) Google Scholar) was used, and torsion angle dynamics (TAD) combined with a simulated annealing algorithm was employed in the calculation. A sequence alignment of the ACPs from M. tuberculosis andM. leprae with typical bacterial ACPs is shown in Fig.1. The proteins have a high degree of similarity, especially around the serine residue where the prosthetic group is attached and extending along helix II. The most notable difference between AcpM and other ACPs that function in the type II fatty acid synthase systems is the carboxyl-terminal extension. A search of current sequence data bases failed to detect any significant similarity between the unique AcpM carboxyl-terminal extension and other sequences. The structure of AcpM from M. tuberculosis was studied by multidimensional NMR spectroscopy. The resonance assignments were obtained from an array of heteronuclear experiments. The backbone assignments were achieved through conventional strategy by using CBCANH and HN(CO)CBCA experiments (25Clore G.M. Gronenborn A.M. Methods Enzymol. 1994; 239: 349-363Crossref PubMed Scopus (250) Google Scholar, 26Grzesiek S. Bax A. J. Biomol. NMR. 1993; 3: 185-204Crossref PubMed Scopus (708) Google Scholar, 27Wong H.C. Mao J. Nguyen J.T. Srinivas S. Li, L., Wu W. Zhang D. Zheng J. Nat. Struct. Biol. 2000; 7: 1178-1184Crossref PubMed Scopus (125) Google Scholar). The side-chain assignments were mainly derived from 13C-HCCH-TOCSY and13C-HCCH-COSY data. The NOE constraints were obtained from both 15N-edited NOESY and15N/13C-edited NOESY spectra. A total of 2043 unique NOE distance restraints were identified; a summary of these constraints is illustrated in Fig. 2. On the basis of resonance assignments using the program CSI (20Wishart D.S. Sykes B.D. J. Biomol. NMR. 1994; 4: 171-180Crossref PubMed Scopus (1893) Google Scholar) and the chemical shift values (mainly the Cα and CO chemical shifts), four helical regions were clearly identified. This result was consistent with the distribution of NOEs; within these regions, helical characteristic medium range, Hαi-HNi+3and Hαi-Hβi+3, as well as strong sequential HNi-HNi+1 NOEs were clearly observed. In the carboxyl-terminal extension, a few helical elements were detected by CSI. However, other than intra-residence and sequential NOE distance constraints, few NOEs were found, which indicates that there was no defined structure in this section of the protein. The structure of the AcpM was determined based on 2043 NOE distance constraints and 48 hydrogen bond restraints derived from NMR measurements. In the final calculation, a total 320 structures were calculated using the program DYANA (24Guntert P. Mumenthaler C. Wuthrich K. J. Mol. Biol. 1997; 273: 283-298Crossref PubMed Scopus (2537) Google Scholar); 20 structures with the lowest target functions were selected and superimposed (Fig. 3). The structural statistics are given in Table I. All experimental NMR constraints were well stratified, and there were no NOE constraint violations of more than 0.35 Å. Although no dihedral torsion angle constraint was applied in the structural calculation, most residues have Φ and Ψ dihedral angles in the most favorable or favorable regions of the Ramachandran plot (Table I). The only residue that is in the disallowed region is Phe-33, which is located in the long loop between two helices. The solution structure was of high precision. The average root mean square deviation from the average structure for backbone heavy atoms (C′, Cα, N) of the first 90 residues was 0.42 Å. Within the four well folded helical regions, the average root mean square deviation for backbone heavy atoms was 0.31 Å.Table IStatistics for AcpM structureNumber of restraintsNOE distance restraints Intra-residue1108 Inter-residueShort range465Medium range340Long range130 Total NOE distances2043 Hydrogen bonds48Residues 3–84Secondary structureAverage r.m.s.d. from the mean structure (Å)1-aThe average root mean square deviation (r.m.s.d.) between the 20 structures of the lowest target functions and the mean coordinates of the protein. Backbone0.420.31 Heavy atoms0.860.68Average r.m.s.d. from experimental distance restraints (Å)0.0040 ± 0.0004Residues in most favorable regions1-bExcluding glycine and proline.67.2%74.3%Residues in additional allowed regions28.1%21.5%Residues in generously allowed regions3.4%4.2%Residues in non-allowed regions1.3%0.0%1-a The average root mean square deviation (r.m.s.d.) between the 20 structures of the lowest target functions and the mean coordinates of the protein.1-b Excluding glycine and proline. Open table in a new tab The structure of AcpM consists of a folded amino-terminal region and a highly flexible and structurally undefined carboxyl terminus. The four α-helices in the structure of AcpM comprise a “right-turn” helical bundle (Fig. 1). Its topology is “square,” as classified by Cohen and co-workers (28Harris N.L. Presnell S.R. Cohen F.E. J. Mol. Biol. 1994; 236: 1356-1368Crossref PubMed Scopus (131) Google Scholar). In antiparallel helix packing, all four helices (I–IV) are connected by “underhand” loops. Helix I (residues Gln-5 to Val-19), II (Asp-40 to Lys-54), and IV (Val-70 to Glu-81) are of approximately equal length with an up-down-down topology. Helix III (Asp-61 to Leu-67) is relatively short compared with the other helices. The length of the loops that connect these four helices is also variable. The loop between helices I and II is the longest and contains about 20 residues. The four-helix bundle is the central architectural feature of the AcpM, and this helical bundle is maintained by hydrophobic interactions between the helices. The helix I contains seven hydrophobic residues. Ile-12, Ile-15, and Val-19 form hydrophobic contacts with Met-44 and Ile-47 (contacts with both Ile-15 and Val-19) from helix II, and residues Ile-8 and Ile-12 are in close contact with Val-74 and Val-73 of helix IV. Helix III, which is the shortest helix, has contacts with Ala-48 in helix II and Tyr-76 in helix IV through residue Leu-64. The three loops that connect the four helices in the ACP domain are all highly ordered. Loop 1, which connects helices I and II, is the longest, and despite the large number of residues, it is not flexible and has a defined structure. Several hydrophobic contacts between the residues within the loop and residues in the helical core keep the loop in close contact with the helical bundle. Notably, the side chains of residues Ile-27 and Pro-29 within the loop have contacts with Ile-16 and Ile-9 in helix I. The loop is further stabilized by an intra-loop salt bridge, Glu-26 and Lys-31. Loop 2 and Loop 3 connect helix III to the main helical bundle, and residues in both loops have contacts to the main helical core. For example, Ile-59 in Loop 2 and Leu-67 in Loop 3 form hydrophobic contacts with Ile-77 and Tyr-76 in helix IV, respectively. Furthermore, Lys-58 forms a salt bridge with Glu-81 in the carboxyl terminus of helix IV. These interactions function to reduce the flexibility of the helix III, the shortest helix in the structure. NMR structures of the E. coli ACP (29Kim Y. Prestegard J.H. Proteins. 1990; 8: 377-385Crossref PubMed Scopus (156) Google Scholar, 30Holak T.A. Kearsley S.K. Kim Y. Prestegard J.H. Biochemistry. 1988; 27: 6135-6142Crossref PubMed Scopus (95) Google Scholar), actinorhodin apoACP (9Crump M.P. Crosby J. Dempsey C.E. Parkinson J.A. Murray M. Hopwood D.A. Simpson T.J. Biochemistry. 1997; 36: 6000-6008Crossref PubMed Scopus (141) Google Scholar) and Bacillus subtilis ACP (7Xu G.Y. Tam A. Lin L. Hixon J. Fritz C.C. Powers R. Structure. 2001; 9: 277-287Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar), as well as the crystal structure of the B. subtilis ACP in the ACP·AcpS complex (8Parris K.D. Lin L. Tam A. Mathew R. Hixon J. Stahl M. Fritz C.C. Seehra J. Somers W.S. Structure. 2000; 8: 883-895Abstract Full Text Full Text PDF PubMed Scopus (200) Google Scholar) have been reported. The secondary structure elements and overall folding of these proteins are very similar. Comparison of the AcpM solution structure with these ACP structures indicates that the structured amino-terminal domain of AcpM possesses the common ACP fold. The superposition of AcpM with each of the known ACP structures over backbone atoms of the helical core is illustrated in Fig.4. The structural homology is obvious and extends to all four helices. Variations in the structures are apparent in the loop regions, although it is not clear whether any of these differences could be attributed to the conditions of data collection and structural refinement. Despite the differences in sequences among the ACPs (Fig. 1), a detailed examination revealed that the interhelical hydrophobic interactions, which stabilize the helical bundle of ACPs, are very similar. For example, in our AcpM structure, Ala-48 in helix II forms hydrophobic contacts with Val-73 in helix IV. In the three other bacterial ACP structures, the residue corresponding to the AcpM Ala-48 in helix II is valine (Val-43 in E. coliACP sequence), whereas the residue in helix IV corresponding to Val-73 is alanine (Ala-68 in E. coli ACP). Therefore, the nature of the alanine-valine contact is the same in all the ACPs. The DSL motif (Asp-40, Ser-41, and Leu-42 in the AcpM sequence) is conserved in all of the ACP sequences, and the 4′-PP prosthetic group is covalently linked via a phosphodiester bond to the serine residue. Closer examinations of all four ACP structures revealed that a high degree of structure homology exists in the region adjacent to the DSL motif. In all four structures, the DSL sequence is present at the amino terminus of helix II (Fig. 4). This strong structural homology of the DSL region in AcpM to other bacterial ACPs provides a rationale for why AcpS and the enzymes of E. coli type II fatty acid synthesis use ACPs from other species as well as the species-specific ACP. The similarities are strongest along helix II, a domain of the protein referred to as the recognition helix, which is responsible for the interaction of ACPs with the enzymes of type II fatty acid synthesis (10Zhang Y.M. Rao M.S. Heath R.J. Price A.C. Olson A.J. Rock C.O. White S.W. J. Biol. Chem. 2001; 276: 8231-8238Abstract Full Text Full Text PDF PubMed Scopus (147) Google Scholar). Increased mobility in the carboxyl-terminal segment is clear from both the relaxation (Fig. 5) and solvent exchange/accessibility data (data not shown), suggesting that the carboxyl-terminal domain of AcpM exists as a “molten domain.” The steady-state heteronuclear 15N[1H] NOE value versus the residue number is shown in Fig. 5. Because the lengths of the N–H bonds are fixed, the15N[1H] NOE values report information about the dynamics of N–H bonds and are used to determine whether a particular amide is in a folded or unfolded region of a protein. The value of the heteronuclear 15N[1H] NOE for folded residues is ∼0.7–1; the NOE value for a flexible loop is ∼0.3–0.5; and the NOE value for the unstructured residues is between −1.0 and 0 (31Fushman D. Najmabadi-Haske T. Cahill S. Zheng J. LeVine III, H. Cowburn D. J. Biol. Chem. 1998; 273: 2835-2843Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). Although we did not detect any defined secondary structural elements in the carboxyl-terminal domain, the structure of this region remains relatively ordered until Glu-96. Past this point, the values of 1H[15N] NOEs are less than zero, indicating a completely melted down random coil. The dynamic study showed the gradual loss of NOEs in this region, which is characteristic for an unfolded structure. Nevertheless, the AcpM NOESY data indicate several close contacts between the carboxyl-terminal domain and helix I in the amino-terminal ACP fold, mainly in the first few residues of the sequence. Specifically, we observed NOE contacts between Thr-4 and Gln-5 in helix I with Gln-89. A special term, “natively unfolded,” is applied to protein domains that have little or no ordered structure under physiological conditions (32Wright P.E. Dyson H.J. J. Mol. Biol. 1999; 293: 321-331Crossref PubMed Scopus (2278) Google Scholar, 33Ma C.J. Tsai B. Sham Y.Y. Kumar S. Nussinov R. Proteins. 2001; 44: 418-427Crossref PubMed Scopus (176) Google Scholar, 34Uversky V.N. Gillespie J.R. Fink A.L. Proteins. 2000; 41: 415-427Crossref PubMed Scopus (1724) Google Scholar); many intrinsically unstructured proteins have been identified and studied (34Uversky V.N. Gillespie J.R. Fink A.L. Proteins. 2000; 41: 415-427Crossref PubMed Scopus (1724) Google Scholar). Unstructured proteins are inherently flexible, and their conformations are readily shaped through interactions with other molecules (32Wright P.E. Dyson H.J. J. Mol. Biol. 1999; 293: 321-331Crossref PubMed Scopus (2278) Google Scholar, 35Xu C.J. Tsai D. Nussinov R. Fold. Des. 1998; 3: R71-R80Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar). The intrinsic plasticity of “natively folded” proteins offers several important advantages in regulatory systems in which the unstructured protein is induced to interact with a number of different partners (32Wright P.E. Dyson H.J. J. Mol. Biol. 1999; 293: 321-331Crossref PubMed Scopus (2278) Google Scholar). Indeed, the requirement for a folding transition upon binding of a disordered protein domain to a target contributes significantly to the specificity of the interaction (32Wright P.E. Dyson H.J. J. Mol. Biol. 1999; 293: 321-331Crossref PubMed Scopus (2278) Google Scholar). The molten down, unfolded structure of the carboxyl-terminal extension of AcpM has the hallmark features of a natively folded domain, suggesting that it plays multiple roles in the function of the AcpM. The mycobacterial type II fatty acid synthase produces very long chain mycolic acids that give rise to the unique cell envelope structure of this group of organisms (36Brennan P.J. Nikaido H. Annu. Rev. Biochem. 1995; 64: 29-63Crossref PubMed Scopus (1524) Google Scholar). Thus, one possible function of the unique AcpM carboxyl-terminal domain is to interact with the variety of long-chain fatty acid intermediates carried by the protein (15Barry III, C.E. Lee R.E. Mdluli K. Sampson A.E. Schroeder B.G. Slayden R.A. Yuan Y. Prog. Lipid Res. 1998; 37: 143-179Crossref PubMed Scopus (442) Google Scholar, 37Kolattukudy P.E. Fernandes N.D. Azad A.K. Fitzmaurice A.M. Sirakova T.D. Mol. Microbiol. 1997; 24: 263-270Crossref PubMed Scopus (161) Google Scholar). More than half the residues in the unstructured AcpM domain are classified as hydrophobic, and the domain may function to sequester from solvent the steadily elongating acyl chain bound to the prosthetic group. In this scenario, the unstructured carboxyl-terminal domain in AcpM would fold to a different extent to accommodate the spectrum of long-chain and very long chain acyl intermediates in mycolic acid biosynthesis. Experimental validation of this idea awaits the development of technologies that allow the synthesis of very long-chain acyl-AcpM. On the other hand, it is also plausible that the carboxyl-terminal extension may be involved in protein-protein interactions. In the type II fatty acid synthase, ACP interacts with a variety of functionally different enzymes (10Zhang Y.M. Rao M.S. Heath R.J. Price A.C. Olson A.J. Rock C.O. White S.W. J. Biol. Chem. 2001; 276: 8231-8238Abstract Full Text Full Text PDF PubMed Scopus (147) Google Scholar). The unstructured domain may allow AcpM to interact specifically with pathway enzymes, perhaps by presenting a slightly different structure for each acyl intermediate and thereby directing the cargo to the appropriate pathway enzyme. The carboxyl-terminal domain may also act as an inhibitor of the interaction of acyl-AcpM with acyltransferases. It is interesting that long-chain acyl-AcpM accumulate in E. coli, suggesting that AcpM is able to productively interact with the enzymes of type II fatty acid synthesis but is not a substrate for the glycerol-phosphate acyltransferases involved in the formation of membrane phospholipids. Thus, the carboxyl terminus may function to prevent the acyl chains destined for mycolic acids from being diverted to membrane phospholipid formation. Clearly, further studies of the AcpM structure(s) in the presence of proteins that it interacts with will shed the light on the role of the carboxyl-terminal domain mycolic acid synthesis. The 4′-PP prosthetic group is covalently attached to the Ser-41 in AcpM. Based on the observation that no NOE was detected between the 4′-PP and ACP residues in B. subtilis ACP, it was suggested that the 4′-PP is readily accessible at the protein surface. Because our15N/13C-labeled AcpM sample had about 60% of the 4′-PP molecules unlabeled, we were able to perform a number of “filtered-edited” experiments (38Zwahlen C. Legault P. Vincent S. Greenblatt J. Konrat R. Kay L.E. J. Am. Chem. Soc. 1997; 119: 6711-6721Crossref Scopus (533) Google Scholar) to study the interaction between 4′-PP and protein. Consistent with the observations reported in previous NMR ACP structures (7Xu G.Y. Tam A. Lin L. Hixon J. Fritz C.C. Powers R. Structure. 2001; 9: 277-287Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar), no inter-NOEs between the prosthetic group and protein itself were detected in our experiments. Our analysis of the differences in the amide proton and nitrogen chemical shift of residues in apoAcpM and AcpM revealed a clearer picture of the prosthetic group dynamics. 15N-HSQC spectrum of the mixture of apo- and holoforms of 15N labeled AcpM showed that majority of the residues in the AcpM have same amide proton and nitrogen chemical shift. However, about a dozen residues had different amide proton and/or nitrogen chemical shifts in the apo- compared with the holoform. The differences were small enough so that we were easily able to assign every residue in the apo form based on the resonance assignment in the holoprotein. No significant chemical shift differences were noted between apoAcpM and AcpM, suggesting that the structures of the two forms are essentially the same. Nevertheless, the observation of the chemical shift differences between the two forms shows that the bound 4′-PP group interacts with the protein. Using the equation provided by Kurt Wuthrich and co-workers (39Pellecchia M. Sebbel P. Hermanns U. Wuthrich K. Glockshuber R. Nat. Struct. Biol. 1999; 6: 336-339Crossref PubMed Scopus (105) Google Scholar) to combine the chemical shift difference of proton and nitrogen shift together, we summarized the chemical shift changes between apoAcpM and AcpM (see Fig. 6). The diameter of the sausage corresponds to the sum of 1H and 15N shift perturbation due to the bound 4′-PP group. The chemical shift perturbations were detected in residues that range from helix I to the loop that connects helix III and helix IV. No perturbations were observed for helix IV and the carboxyl-terminal coil. Large perturbations were observed for Phe-33, Asp-40, Ser-41, Leu-42, Thr-51, Glu-52, Ala-65 and Gly 66. These residues, with the exception of Thr-51 and Glu-52, are structurally close to Ser-41, which is covalently linked to the 4′-PP group. Thr-51 and Glu-52, located in the carboxyl terminus of the helix II, were also perturbed, indicating transient interaction with the prosthetic group. Small perturbations were also found in the loop region that connects helix II with helix III (Fig.6). Together, those residues form an extended hydrophobic region around Ser-41. The chemical shift of a nucleus is sensitive to changes in the local environment, including aromatic ring-current effects, peptide bond anisotropy, electrostatic interactions, and hydrogen bonding. When a ligand binds to a protein, the interactions between them cause changes in the environment of the nuclei at the interfaces, resulting in chemical shift changes. The analysis of the AcpM chemical shift data clearly points to the 4′-PP group spending part of the time bound to the hydrophobic pocket. It is likely that the attached 4′-PP group experiences a rapid exchange between two states, one that is bound to the hydrophobic pocket and another that is solvent-exposed. These two states are consistent with the dual function of ACPs, which need to both load and unload cargo (acyl intermediates) into the active sites of the type II enzymes while protecting the cargo from solvent when shuttling the fatty acyl intermediates between the enzymes of type II fatty acid synthesis. We thank Tuan Tran for technical assistance and Dr. Weixing Zhang for NMR technical support.