The apicomplexan Cryptosporidium parvum is an intestinal parasite that affects healthy humans and animals, and causes an unrelenting infection in immunocompromised individuals such as AIDS patients. We report the complete genome sequence of C. parvum, type II isolate. Genome analysis identifies extremely streamlined metabolic pathways and a reliance on the host for nutrients. In contrast to Plasmodium and Toxoplasma, the parasite lacks an apicoplast and its genome, and possesses a degenerate mitochondrion that has lost its genome. Several novel classes of cell-surface and secreted proteins with a potential role in host interactions and pathogenesis were also detected. Elucidation of the core metabolism, including enzymes with high similarities to bacterial and plant counterparts, opens new avenues for drug development.

incubated with 20 ml Sepharose-immobilized monoclonal anti-HA antibodies (Covance). Beads were washed with buffer T without BSA before elution. DSP-induced crosslinks in eluted proteins were thiol-cleaved before separation by 8-16% SDS-PAGE and detection by Sypro Ruby (Bio-Rad). Quinone analysesLipid extractions and quinone detection were performed as described 19 . Hydrogenosomes (5.5 mg protein) were extracted and resuspended in 150 ml 9:1 methanol/ethanol, of which 50 ml was injected onto an HPLC system linked to an ECD. Sequence analysesAccession numbers for sequences used to reconstruct NuoF and NuoE phylogenies are listed in Supplementary Tables 2 and 3. NuoF sequences were aligned with CLUSTALX. NuoE sequences were aligned with Wisconsin Package Version 10.2 programs (Genetics Computer Group). A profile hidden Markov model (HMM) was built from Escherichia coli, Neurospora crassa, Bos taurus, Paracoccus denitrificans and Thermus thermophilus sequences with HMMBUILD. Additional sequences were aligned to the profile with HMMALIGN. Both alignments were edited to remove C-and N-terminal extensions. Analyses of NuoF and NuoE evolution were performed with MRBAYES 30 with the JTT amino-acid substitution model and with two Markov chains Monte Carlo. Chains were run for 100,000 generations, with sampling every 50 generations. The first 5,000 generations were discarded as burn-in. Consensus trees satisfying the more than 50 majority rule were drawn with Treeview, and probabilities of branch partitions were calculated.