Proceedings of Bioc 462a Honors

Abstract
Acetyl-CoA carboxylase (ACC) is an essential enzyme in all living organisms, providing catalysis for the rate limiting step in the synthesis of fatty acids. Several structures have been determined for specific subunits and domains of this enzyme, but no structural information exists to date for an ACC in Aedes aegypti mosquitoes or any invertebrate species. This paper seeks to draw information from known structures to make conclusions about the unknown A. aegypti ACC structure and function. Several bioinformatics databases were used to determine the sequence homology between the mosquito ACC and those with known structures. It was found that extensive homology exists, and it was concluded that these structures give a good representation of the unknown structure. Specifically, structures of the biotin carboxylase (BC) domain of ACC are studied and used to extrapolate information about probable characteristics of the mosquito enzyme. There is a high degree of conservation in the active site of BC and also an allosteric binding site for soraphen A. Because of the high sequence homology across ACC BC species, a unique pharmaceutical target for vector control is questionable.

Introduction Acetyl-CoA carboxylase (ACC), an enzyme found in all living organisms, catalyzes the first and rate limiting step in the synthesis of long-chain fatty acids (Waldrop et al., 1994). In this reaction, acetyl-CoA is converted into malonyl-CoA in a biotin-dependent manner. The malonyl-CoA can then be elongated to make palmitate by the enzyme fatty acid sythetase (Wakil et al., 1983). In mammals, a second isozyme of acetyl-CoA carboxylase is involved in the regulation of fatty acid oxidation, in which the malonyl-CoA product is a potent inhibitor in the lipid oxidation pathway (Tong, 2005).
Because of its important role in the synthesis and regulation of fatty acids, ACC is an essential and highly conserved enzyme. In prokaryotes, ACC exists in multiple subunits, including biotin carboxylase (BC), biotin carboxyl carrier protein (BCCP), and carboxyl transferase (CT). In most eukaryotes, ACC is a large multi-domain protein whose catalytic domains have a high sequence similarity with the bacterial subunits (Tong, 2005). The BC domain carboxylizes the biotin moiety on the BCCP domain in an Mg₂∙ATP dependent reaction. The carbon is then transferred to acetyl-CoA by the CT domain to finally form malonyl-CoA (Wakil et al., 1983)(Figure 1).

Figure 1. Reaction catalyzed by ACC. The first step shows the reaction carried out in the BC domain/subunit. Not shown is the first step in which ATP and carbonate are bound to form a carboxyphosphate intermediate. The intermediate is attacked by biotin so that carbon can be transferred to the next active site. The second reaction occurs in the CT domain. In this step acetyl-CoA is carboxylized to form malonyl-CoA (Thoden et al., 2000). (Larger View).

The BC domain belongs to a large family of proteins characterized by an ATP-grasp fold, for which the overall chemical transformation is the formation of a carbon-nitrogen bond (Thoden et al., 2000). In BC, ATP and bicarbonate are first bound to form a carboxyphosphate intermediate. This intermediate decomposes into inorganic phosphate and carbon dioxide. The phosphate helps to deprotonate the N1 atom of biotin which is covalently attached to the BCCP domain. The biotin can then attack carbon dioxide to produce carboxybiotin (Tong, 2005). The active sites of BC and CT are 7Å apart allowing the BCCP domain to undergo a large conformational change, transferring carboxybiotin between these sites for the second part of the ACC reaction (Waldrop et al., 1994).


The current emergence of an obesity epidemic, as well as similar metabolic conditions, has generated a great deal of interest in the enzymes involved in fatty acid synthesis and oxidation, including of course ACC. Abu-Elheiga et al. used ACC2 deficient mice mutants to show that controlling this enzyme produces mice that have a normal life span, a higher fatty acid oxidation rate, and lower amounts of body fat. Without the gene encoding ACC2, these mice have continuous fatty acid oxidation because of malonyl-CoA depletion, despite consuming more food. This offers them protection against diabetes and obesity even in the presence of a high carbohydrate/high fat diet (Abu-Elheiga 2001). As the world human population consumes an increasing amount of carbohydrates and fat, a drug aimed at decreasing ACC2 expression could prove incredibly monumental.
Extensive research has been done on the mammalian ACC because of its potential as a target against metabolic disorders, but so far no one has considered the possibility of targeting the ACC of hematophagous (blood-feeding) invertebrates. Hematophagous invertebrates, particularly mosquitoes, are very important to human disease because of their role in spreading vector borne pathogens. There is relatively little lipid in the mosquito diet so lipid synthesis is vital for egg production and maintaining maternal reserves (Briegel 1990). Lipids are synthesized from both the sugar meal and the blood meal taken by the mosquito. In the process of consuming up to three times or more their own body weight in blood, female mosquitoes often ingest or transmit pathogenic diseases (Marquardt, 2004). Therefore, controlling lipid synthesis could be a novel approach in vector control, which ultimately could prevent the spread of devastating pathogens.

A thorough understanding of an enzyme’s structure sheds light on the mechanisms that could potentially be exploited for pharmaceutical purposes. Here, the knowledge obtained from previous studies of known ACC biotin carboxylase structures is used to extrapolate information about the Aedes aegypti, or yellow fever mosquito ACC structure which has yet to be determined. Because of the high degree of homology between all ACC protein sequences, a fair degree of homology between structures can be assumed. Structural information can be used to determine whether unique targets for vector control are possible.

Methods and Results

Methods for determining first order sequence homology
Vectorbase.org is an online bioinformatics resource for common invertebrate disease vectors including A. aegypti mosquitoes (Lawson et al., 2007). Its BLAST (Basic Local Alignment Search Tool) feature allows query proteins to be aligned with complete or partially complete vector genomes (Altschul et al., 1997, Nene et al., 2007). In order to determine whether an ACC gene sequence exists in the known A. aegypti genome, an animal ACC was sought for alignment using BLAST. When the human ACC2 (2458 amino acids) was used for sequence comparison, one mosquito gene sequence (2529 amino acids) had an E-value of 0.0. The E-value is the number of hits with a better match one would expect to get by chance. This sequence had a 60.4% identity match with the human ACC2 gene. Human ACC1 is more similar to the same gene hit at 64.8%.
Figure 2 shows a partial sequence alignment (285 amino acids) between human ACC2 and A. aegypti ACC. Human ACC2 was chosen to show alignment because the structure of its BC domain has been ascertained. The region shown includes the N-terminus of the biotin carboxylase domain. Ser222 is highlighted because of its importance in ACC regulation, discussed below. A glycine rich stretch is also highlighted.

Figure 2. Partial sequence alignment of human ACC2 and A. aegypti putative ACC. BLAST was used on vectorbase.org to make this sequence alignment (Lawson et al., 2007, Altschul et al., 1997). The program identified a 60.4% identity. The human sequence is in black, the mosquito in blue, and the shared residues in pink. Similar but not identical residues are marked with a +. Serine222 of human ACC2 is highlighted as well as an important stretch of glycine residues. (Larger View)

The difference in sequence homology could be important in revealing the main function of ACC in A. aegypti. As discussed in the introduction, two isozymes of ACC have been discovered in mammals. ACC1 is localized in the cytosol of the cell and is an essential player in the synthesis of fatty acids. ACC2, encoded by a different gene, is localized in the mitochondrial membrane where it can inhibit important proteins involved in fatty acid oxidation. The two genes have considerable identity except near the N-terminus. This region is hypothesized to be involved in anchoring the enzyme to the mitochondrial membrane in ACC2 (Munday, 2002). Currently, no research has been done to determine the role or intracellular location of ACC in mosquitoes, but the BLAST results suggest that the function is most similar to mammalian ACC1.

The European Bioinformatics Institute (EBI) has a valuable resource, InterProScan, that reveals detailed information about protein families and domains (Zdobnov et al., 2001). It was used to help confirm the identity of the A. aegypti ACC sequence and to elucidate the important functional groups. InterProScan identified the major domains found in all studied ACCs, including the BC and CT domains. It also recognized an ATP grasp fold in the BC domain and a biotin binding site in a region between the two domains, probably an uncharacterized BCCP domain. Interestingly, mosquito and human scans came back with nearly identical results (Figure 3).

Figure 3. InterProScan results. An InterProScan was done on both the A. aegypti ACC (A) and human ACC2 (B). Results show important domains identified by the scan, colored regions. Shown in this way, the similarity between species is significant (Zdobnov et al., 2001).
(Larger View)

The start and end points of the BC region were noted for secondary and tertiary structure comparison, discussed below. Sequence comparisons of BCs were also made using NCBI’s BLAST engine (Altschul et al., 1997). One available function of the BLAST engine is a tool that aligns two specific gene or protein sequences. BCs with known protein structures were aligned with A. aegypti BC and each other. Results for yeast, human ACC2, and E. coli BCs are reported in Figure 4. The primary structure homology of ACC biotin carboxylases demonstrates that this region is, predictably, highly conserved across eukaryotic and even prokaryotic species.

Figure 4. Amino acid conservation of the BC domain of multiple species. Numbers are percent identity. They were determined using NCBI’s align 2 sequences BLAST tool (Tong, 2005, Altschul et al., 1997). (Larger View)

Important features of known biotin carboxylase structures
This review will focus on known structural information for the yeast BC domain, the human ACC2 BC domain, and the E. coli BC subunit. Although BC exists as one part of a multi-domain or multi-subunit enzyme, it can itself be separated into several domains. The A and C domains form a cylindrical structure that constitutes the main body of the enzyme and also contains the active site. The B domain creates a lid over the active site and also contributes to the molecular interactions between substrates and enzyme. In the structures of yeast and human BC, an AB linker between the A and B domains is also identified (Tong, 2005) (Figure 5A).

Figure 5. Important structural features of BC. All images rendered using PyMOL. (A) Yeast ACC BC shown in cartoons. The A domain is shown in orange, B in green, C in blue, and the AB linker in magenta. Accession code 1W93 (Shen et al., 2004). (B) The binding site of soraphen A from yeast BC. The active site is shown in surface mode and the ligand in a stick model. Assession code 1W96 (Shen et al., 2004). (C) The binding site of ATP from E. coli BC mutant E288K. Similarlarly, The active site is shown in surface mode and the substrate in a stick model. The surface is rendered slightly transparent so that binding can be fully visualized. Accession code 1DV2 (Thoden et al., 1999).
(Larger View)

The B domain is thought to undergo a large conformational change during catalysis. In prokaryotic BC, the B domain assumes a significantly more open position in the absence of ATP. In the eukaryotic structures, such a significant change is not noticed, but instead the B domain remains closed even in the absence of ATP (Tong, 2005). The flexibility of the B domain of E. coli could be accounted for by the high degree of disorder in the amino acids present. In a stretch of approximately 70 residues, 15% are glycine residues, including five in sequence (Thoden et al., 2000).

A very important feature of BC is the ATP-grasp fold. BC belongs to a large family of proteins that contain this tertiary structure. Studies of BC were actually very important in shedding light on the reaction mechanisms characteristic of these proteins. In all of these proteins, ATP must first bind to a cleft formed by two major structural elements including two beta strands and an alpha helix (Galperin and Koonin, 1997). Some sort of carbon donating molecule must also be bound to form the carboxyphosphate intermediate. The carboxyphosphate intermediate then reacts with the biotin moiety attached to the BCCP domain of ACC (Tong et al., 2005).
An E288K mutant E. coli BC subunit, in which glutamate 288 was changed to a lysine, was able to form a stable complex with ATP in a crystal structure (Figure 5C). ATP binds to the active site at the interface between the B and C domains. Hydrogen bonds form between the enzyme and the adenine base of ATP. The phosphates interact with a glycine rich loop and several positively charged residues. Magnesium ions, necessary for the reaction to occur, are not shown in the structure because it is hypothesized that Glu288 coordinates these cations. A glutamate appears in this region in the other species, including A. aegypti. In the bacterial mutant, the lysine residue takes the place of magnesium. An interesting experiment to try on A. aegypti ACC would be to add a metal chelator to see its effect on lipid synthesis. Biotin, which is bound to the BCCP subunit, also interacts with this binding pocket. Again, charged residues are situated to interact with biotin (Thoden et al., 2000).
The yeast BC structure was solved alone and in complex with soraphen A, a potent biotin carboxylase inhibitor. Soraphen A binds to an allosteric site opposite ATP on the cylindrical structure formed by the A and C domains. It interacts with several residues in this region, and these residues are highly conserved in eukaryotic species. E. coli BC does not bind soraphen A tightly, probably due to steric hindrance caused by a nearby alpha helix. A novel mode of inhibition was proposed for soraphen A. Soraphen A binds near the putative dimerization site of yeast BC as determined by the known dimerization site of bacterial BC. Soraphen A may inhibit dimer formation, suggesting that dimerization is required for active enzymatic function (Shen et al., 2004)(Figure 5B).
E. coli BC exists as a dimer in solution, with each subunit situated so that the B domains are on opposite edges of the dimer. Each subunit contains its own active site, and no cooperativity is observed in substrate binding (Waldrop et al., 1994). Yeast and human BCs do not exist as dimers in solution, and it is not known if dimerization occurs in the active enzyme. The residues constituting the hypothetical dimerization site are relatively dissimilar to the residues in the bacterial BC dimer interface (Shen et al., 2004). If one were to accept the theory that soraphen A inhibits dimerization in eukaryotic ACC BC domains, soraphen A injection in mosquitoes could tell whether dimerization is important to A. aegypti ACC activity.

Important structural differences between eukaryotic BC structures
Some important structural differences between prokaryotic and eukaryotic BCs have already been discussed above. Because A. aegypti is a eukaryotic species, the discussion of differences between the eukaryotic structures is much more relevant and provides a template for gleaning information about important structural motifs in the unknown structure. The structure of human ACC2 BC has only recently been determined so this information is relatively new and exciting.
Human and yeast BC domains have a high degree of sequence homology, and their overall architecture is quite similar. The B domain may retain a more closed structure in human BC, but this could be an artifact of crystal packing. Another major structural difference is seen in the N-termini of these structures. There is relatively little sequence homology in this region, and the human N-terminus has a higher degree of disorder. The disorder of the human BC N-terminus could be due to a high concentration of basic residues. This could allow for a conformational change when Ser222 is phosphorylated by AMPK, thus explaining how AMPK regulates BC activity. AMPK phosphorylation is an important mechanism for inhibiting mammalian ACC activity. Yeast BC lacks a serine in its N-terminus and lacks the regulatory mechanism inferred by AMPK (Cho et al., 2007).

The structure of the putative soraphen A binding site of human BC is very similar to the structure determined for yeast BC in complex with soraphen A. Trp681 and Met594 have unfavorable side chain conformations near the putative binding site however, and probably undergo a significant conformational change for binding to occur. The binding affinity between human BC and soraphen A is still high due to extensive interactions between the inhibitor and the binding pocket. Whether soraphen A disrupts dimer formation of the human BC, as is hypothesized for the yeast BC, is unknown (Cho et al., 2007).

Using known models to predict mosquito BC structure and function
The fact that only one gene homologous to ACC was found in the A. aegypti genome is itself telling about the function of this enzyme. Only one isoform of ACC is found in yeast and bacteria, but two isoforms exist in mammals. The two isoforms of mammals, although encoded by different genes, are nearly identical (Munday, 2002). If only one isoform is found in the A. aegypti genome, then this suggests that the second isoform had not yet evolved in this invertebrate species. A. aegypti ACC is most similar to human ACC1, the more primitive form of ACC. This provides strong evidence that the ACC gene found in the A. aegypti genome is responsible for fatty acid synthesis and not fatty acid oxidation.

Figure 6A shows several visual representations of the degree of homology between A. aegypti BC and known BC structures. The regions blacked out show residues that are neither identical nor similar. Transitioning from human, to yeast, to bacteria, shows an increased number of blacked out regions. Close examination reveals that residues blacked out in one structure may not be blacked out in the next. Because even E. coli BC, with only about 30% identity with eukaryotic BCs, has a similar structure, A. aegypti BC very likely possesses the same folding motifs as the structures shown.

Figure 6. Visualization of structural conservation. All images rendered using PyMOL. (A) Structures with decreasing similarity to A. aegypti BC. Structures are shown with cartoons with cylindrical helices. In order, human ACC2 accession code 2HJW (Cho et al., 2007), yeast accession code 1W93, E.coli BC subunit accession code 2GPW (Shen et al., 2006). (B) Visualization of structure conservation of BC secondary structures. Human ACC2 BC was used for all, accession code 2HJW. Structures shown with spheres. From left to right, helices, sheets, and loops. (Larger View)

Figure 6B shows the similarity of specific secondary structures between A. aegypti BC and human ACC2 BC. In this view, it appears that most of the difference occurs in the helices and loops. Little difference is observed in the beta sheets, emphasizing that buried residues are less susceptible to change.

Both Figure 6 and the sequence alignment in Figure 2 emphasize that when A. aegypti BC is compared to human ACC2 BC, a higher degree of variation in the N-terminus of the protein is observed relative to the rest of the protein. As mentioned above, the N-terminus is the region of highest variation between human ACC1 and human ACC2 (Brownsey et al., 2006). This is probably the main explanation for why A. aegypti ACC is more similar to human ACC1.
Even though differences are observed in the A. aegypti N-terminus, Ser222 is conserved. This residue is important in gene regulation by the mechanism of AMPK phosphorylation. AMPK phophorylation is an important regulator of ACC inactivation in mammalian ACC. It does not regulate ACC in either yeast or bacteria because there is no AMPK activity in these species (Brownsey et al. 2006). As expected, these species lack the serine phosphorylation site in their N-termini. Unlike in yeast and bacteria, AMPK activity does exist in invertebrates so it is likely that this is one mechanism of ACC control in mosquitoes (Hardie and Pan, 2002). Whether mosquito ACC is really controlled in this manner would need to be further tested to make more accurate conclusions than those based on sequence homology.

Another major site of variation among BCs is the B domain. The extent of the closed position could affect binding affinity for ATP, carbonate ions, and/or biotin. However, from the sequence alignment results, it was found that only 7 residues differ from human BC B domain, and only 12 from yeast BC. Of course, more residues differ from the E. coli BC B domain, including fewer glycines and one less glycine in the glycine stretch. These results suggest that A. aegypti BC retains a closed position like its eukaryotic counterparts. Residues important in interacting with substrates are conserved.

Finally, the soraphen A binding site was examined for sequence homology. Looking at the sequence alignment of A. aegypti ACC BC and yeast ACC BC, all but 2 residues involved with binding soraphen A are conserved. A binding assay would be required to determine whether these changes significantly change the affinity for the inhibitor.

From the high degree of homology between A. aegypti ACC BC and BCs with known structures, it can be concluded that the structures are very similar. This suggests that there are few unique regions in the mosquito ACC. In the discussion below, these results will be examined with respect to their relevance to vector control.

Discussion
A high degree of sequence and structure similarity across ACCs in all species highlights the fact that the function and the modes of regulation are also similar. Looking for motifs unique to mosquitoes, then, may seem like a futile pastime. However, very small differences, even differences in one residue could significantly alter the mechanisms involved in regulation. Broad analysis was able to pick out these discrete variations, but more specific studies are needed to analyze their significance.

The fact that such a high degree of homology exists argues that ACC is a critical enzyme in mosquitoes as it is in other species. Knockdown of ACC in yeast essentially kills the organism because it is unable to synthesize lipids which are essential for cell function (Tong, 2005). Knockdown of ACC in mammals could actually be beneficial at some level as shown in mice mutation experiments in obesity and metabolic disorder studies. This difference is due to the fact that mammals ingest a large amount of lipids in their meals, increasingly so in the modern age. (Abu-Elheiga et al. 2001).

Mosquitoes obtain some lipids from a blood meal and also from their larval diet. Experiments done by Zhou et al. used radioactive labeling to determine the composition of nutrients in the eggs and maternal reserves of the mosquito and the source of these nutrients. Much of the lipid comes from the larval diet, even in subsequent gonotrophic (egg laying) cycles, but lipids are replenished after a mosquito has fed on blood and sugar (Zhou et al., 2004a, Zhou et al., 2004b). Most of this lipid is probably synthesized from sugars and amino acids by means of ACC and the lipid synthesis route. Hypothetically, knocking down the ACC gene would have a significant phenotype. Phenotypes might include decreased fat reserves after eggs are laid, decreased egg amount or size, shorter life span, or reduced egg viability. RNA interference is one method for testing this hypothesis. Injections or meals supplemented with ACC inhibitors such as soraphen A might also be interesting.

Other experiments could be done to determine the function and modes of regulation of A. aegypti ACC without determining a crystal structure. Expression profiling of different tissues could provide information about ACC’s location and role in the mosquito just as the location of ACC1 and ACC2 helped identified their respective functions. Substrates, inhibitors, and regulators can be tested in vivo or in vitro to characterize gene function.

If it is found that ACC is indeed critical for mosquito reproduction and/or survival, then distinct mechanisms involved in its control could be used to create novel drugs for vector control. Even if a significantly distinct mechanism is not found, studies of the benefits of ACC knockout in mammals could be combined with vector control studies to make a very unique and advantageous control system.

Acknowledgements
I would like to thank Dr. Jorge Zamora for his guidance throughout this project and the guidance I am sure he will provide in seeing these studies through.

References
Abu-Elheiga L, Matzuk MM, Abo-Hashema KAH, Wakil SJ. (2001) Continuous fatty acid oxidation and reduced fat storage in mice lacking acetyl-CoA carboxylase 2. Science 291, 2613-2616.

Altschul, Stephen F., Thomas L. Madden, Alejandro A. Schaffer, Jinghui Zhang, Zheng Zhang, Webb Miller, and David J. Lipman (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402

Artymiuk P. J., Poirrette A. R., Rice D. W. and Willett P. (1996) Biotin carboxylase comes into the fold. Nat. Struct. Biol. 3, 128–132

Briegel, H., (1990) Metabolic relationship between female body size,
reserves, and fecundity of Aedes aegypti. Journal of Insect Physiology
36, 165–172

Brownsey, R.W., Boone A.N., Elliott J.E., Kulpa J.E. and Lee W.M. (2006) Regulation of acetyl-CoA carboxylase. Biochem Soc Trans. Apr;34(Pt 2):223-7.

Cho YS, Lee JI, Shin D, Kim HT, Cheon YH, Seo CI, Kim YE, Hyun YL, Lee YS, Sugiyama K, Park SY, Ro S, Cho JM, Lee TG, Heo YS. (2007) Crystal structure of the biotin carboxylase domain of human acetyl-CoA carboxylase 2. Proteins [Epub ahead of print]

Galperin MY, Koonin EV. (1997) A diverse superfamily of enzymes with ATP-dependent carboxylase-amine/thiol ligase activity. Protein Sci 6, 2639-2643

Hardie DG, Pan DA. (2002) Regulation of fatty acid synthesis and oxidation by the AMP-activated protein kinase. Biochem Soc Trans. 6,1064-70.

Lawson D, Arensburger P, Atkinson P, Besansky NJ, Bruggner RV, Butler R, Campbell KS, Christophides GK, Christley S, Dialynas E, Emmert D, Hammond M, Hill CA, Kennedy RC, Lobo NF, MacCallum MR, Madey G, Megy K, Redmond S, Russo S, Severson DW, Stinson EO, Topalis P, Zdobnov EM, Birney E, Gelbart WM, Kafatos FC, Louis C, Collins F., (2007) VectorBase: A home for invertebrate vectors of human pathogens. Nucleic Acids Res. 35, D503-5

Marquardt, W. H. (Editor) (2004) Biology of Disease Vectors, Second Edition. Academic Press: Burlington, MA. 15-30.

Munday MR. (2002) Regulation of mammalian acetyl-CoA carboxylase. Biochem Soc Trans 30, 1059-1064.

Nene V, Wortman JR, Lawson D, Haas B, Kodira C, Tu ZJ, Loftus B, Xi Z, Megy K, Grabherr M, Ren Q, Zdobnov EM, Lobo NF, Campbell KS, Brown SE, Bonaldo MF, Zhu J, Sinkins SP, Hogenkamp DG, Amedeo P, Arensburger P, Atkinson PW, Bidwell S, Biedler J, Birney E, Bruggner RV, Costas J, Coy MR, Crabtree J, Crawford M, Debruyn B, Decaprio D, Eiglmeier K, Eisenstadt E, El-Dorry H, Gelbart WM, Gomes SL, Hammond M, Hannick LI, Hogan JR, Holmes MH, Jaffe D, Johnston JS, Kennedy RC, Koo H, Kravitz S, Kriventseva EV, Kulp D, Labutti K, Lee E, Li S, Lovin DD, Mao C, Mauceli E, Menck CF, Miller JR, Montgomery P, Mori A, Nascimento AL, Naveira HF, Nusbaum C, O'leary S, Orvis J, Pertea M, Quesneville H, Reidenbach KR, Rogers YH, Roth CW, Schneider JR, Schatz M, Shumway M, Stanke M, Stinson EO, Tubio JM, Vanzee JP, Verjovski-Almeida S, Werner D, White O, Wyder S, Zeng Q, Zhao Q, Zhao Y, Hill CA, Raikhel AS, Soares MB, Knudson DL, Lee NH, Galagan J, Salzberg SL, Paulsen IT, Dimopoulos G, Collins FH, Birren B, Fraser-Liggett CM, Severson DW., (2007) Genome sequence of Aedes aegypti, a major arbovirus vector.
Science. 316(5832),1718-23

Shen Y, Volrath SL, Weatherly SC, Elich TD, Tong L. (2004) A mechanism for the potent inhibition of eukaryotic acetyl-coenzyme A carboxylase by soraphen A, a macrocyclic polyketide natural product. Mol Cell 16, 881-891

Shen, Y., Chou, C.Y., Chang, G.G., Tong, L. (2006) Is dimerization required for the catalytic activity of bacterial biotin carboxylase? Mol.Cell 22, 807-818

Thoden JB, Blanchard CZ, Holden HM, Waldrop GL. (2000) Movement of the biotin carboxylase B-domain as a result of ATP binding. J Biol Chem 275, 16183-16190.

Tong L. (2005) Acetyl-coenzyme A carboxylase: crucial metabolic enzyme and attractive target for drug discovery. Cell Mol Life Sci 62, 1784-1803

Wakil SJ, Stoops JK, Joshi VC. (1983) Fatty acid synthesis and its regulation. Annu Rev Biochem 52, 537-579

Waldrop GL, Rayment I, Holden HM. (1994) Three-dimensional structure of the biotin carboxylase subunit of acetyl-CoA carboxylase. Biochemistry 33, 10249-10256

Zdobnov E.M. and Apweiler R. (2001) InterProScan - an integration platform for the signature-recognition methods in InterPro. Bioinformatics17(9), 847-8.

Zhou, G., Flowers, M., Friedrich, K., Horton, J., Pennington, J., Wells, M.A. (2004a) Metabolic fate of [14C]-labeled meal protein amino acids in Aedes aegypti mosquitoes. Journal of Insect Physiology 50, 337–349.

Zhou, G., Pennington, J., Wells, M.A. (2004b) Utilization of pre-existing energy stores of female Aedes aegypti mosquitoes during the first gonotrophic cycle. Insect Biochemistry and Molecular Biology 34, 919–925