Acetyltransferase

551 SNPs were found as genome-wide significant (p < 5E-8) in a recent mega-meta analysis of PD genome-wide association studies [7 (link)]. We defined SNPs in a range of 100 kb of a gene to be within the primary regulatory region of the gene, in which 98 % of cis-acting eQTLs are likely to occur [31 (link)]. Using this criterion, we used Annovar [32 (link)] for the annotation of 166 genes identified as potentially under the regulatory control of these 551 genome-wide significant SNPs. These 166 genes were evaluated in terms of differential RNA or protein abundance in our study (see Table 2 and Additional file 7: Table S7).Table 2

Significant genes with evidence from GWAS analysis

Mega-meta GWAS locia	Implicated gene	Description	Additional evidence	Potential PD-relevant biological functionsb
No rs# available (chr4:816756) intronic	CPLX1	complexin 1	Proteomics	Synaptic vesicle exocytosis
rs2263418 (chr12:40582993) upstream	SLC2A13	solute carrier family 2 (facilitated glucose transporter), member 13	Proteomics	N/A
rs356182 (chr4:90626111) downstream	SNCA	synuclein, alpha (non A4 component of amyloid precursor)	Proteomics	Presynaptic signaling and membrane trafficking
rs8118008 (chr20:3168166) downstream	SLC4A11	solute carrier family 4, sodium borate transporter, member 11	RNA-Seq	N/A
rs4889620 (chr16:31131174) intronic	KAT8	K(lysine) acetyltransferase 8	RNA-Seq	Histone acetyltransferase activity, transcription factor binding
rs6812193 (chr4:77198986) intronic	FAM47E	family with sequence similarity 47, member E	RNA-Seq	Transcription cofactor activity
rs823118 (chr1:205723572) upstream	NUCKS1	nuclear casein kinase and cyclin-dependent kinase substrate 1	RNA-Seq	N/A
rs1375131 (chr2:135954797) UTR3	ZRANB3	zinc finger, RAN-binding domain containing 3	RNA-Seq	DNA annealing helicase and endonuclease activities
rs11724635 (chr4:15737101) upstream	CD38	CD38 molecule	RNA-Seq	Signal transduction, calcium signaling
rs34195153 (chr1:154913723) downstream	PYGO2	pygopus family PHD finger 2	RNA-Seq	Signal transduction

^aSince some genes have multiple SNPs within 100 kb of their start and end sites showing association to PD status at p-values < 5E-8 [7 (link)], only the SNP with the lowest p-value is shown for each implicated gene. SNP coordinates are based on the hg19 (GRCh37.p13) human reference. SNP annotations: downstream/outstream = outside the gene boundaries, within 100 kb from the start or end site of the gene, intronic = located in the gene intron, UTR3 = located in the 3′-UTR of the gene

^bBased on gene information from GeneCards (http://www.genecards.org/), accessed in January 2015

Primers wbdR attB Fw (5′ggggacaagtttgtacaaaaaagcaggcttcATGA ATTTGTATGGTATTTTTGGT3′) and wbdR attB Rv (5′ggggaccactttgtacaagaaagctgggtcTTAAATAGATGTTGGCGA TCTT3′) were designed based on the sequence of E. coli O157:H7¹ and according to Gateway Cloning Technology^® (Invitrogen, Barcelona, Spain) instructions. Both primers were used to amplify wbdR from the start to the stop codon using E. coli O157:H7 DNA as template, and the resulting product was cloned into pDONR221 (Invitrogen) by site-specific recombination, generating pYRI-5. Then, wbdR from pYRI-5 was transferred by site-specific recombination to pRH001 (Hallez et al., 2007 (link)). The new plasmid (pYRI-6) was introduced in Ba-parental (B. abortus 2308W) by conjugation (Conde-Álvarez et al., 2006 (link)) to obtain Ba-pwbdR (Supplementary Table S1).
To obtain a stable B. abortus-wbdR construct, gene wbdR with the 300 bp upstream region containing its promoter was amplified from E. coli O157:H7 using primers wbdR Fw: 5′TTCCCCGGGGGAgaagttcgccacagtaaatcgaa3′ and wbdR Rv: 5′TTCCCCGGGGGAttaaatagatgttggcgatctt3′, and cloned in pGEM-T Easy^® (Promega, Madison, WI, United States) to obtain pYRI-21. The construction was verified by sequencing. Then, the EcoRI fragment of pYRI-21 containing wbdR and its promoter was subcloned into the corresponding site of pUC18R6KT-miniTn7TKm (Llobet et al., 2009 (link)) to obtain pYRI-27 (pUC18R6KT-miniTn7T-Km-PwbdR). The miniTn7 vector carrying wbdR with its own promoter was inserted into Ba-parental chromosome II by the method of Choi et al. (2005) (link) and Choi and Schweizer (2006) (link) modified as follows. First pYRI-27 was introduced in E. coli S17.1 λpir and then transferred to Brucella using an E. coli S17.1 λpir (pYRI-27)-E. coli HB101 (pRK2013)-E. coli SM10 λpir (pTNS2)-Ba-parental four-parental mating. The resulting Ba::Tn7wbdRKm^R construct was examined by PCR for the correct insertion and orientation of miniTn7 between genes glmS and recG using the following primers: (i) GlmS_B (5′-GTCCTTATGGGAACGGACGT-3′) and Ptn7-R (5′-CACAGCATAACTGGACTGATT-3′) for insertion downstream glmS; (ii) Ptn7-L (5′-ATTAGCTTACGACGCTACACCC-3′) and RecG (5′-TATATTCTGGCGAGCGATCC-3′) for insertion downstream recG; and (iii) GlmS_B and RecG that only amplify the intergenic region in the absence of the mini-Tn7. The presence of only one copy of the miniTn7 was determined by Southern-blot and sequencing.
To obtain a wbdR construct with no Km resistance (Ba::Tn7wbdR), a non-polar kmR mutant of Ba::Tn7wbdRKm^R was constructed by overlapping PCR using the Km cassette of pUC18R6KT-miniTn7TKm as template. Primers kmR-F1 (5′-AGGAAGCGGAACACGTAGAA-3′) and kmR-R2 (5′-AATCATGCGAAACGATCCTC-3′) amplified a 318-bp fragment including 312-bp upstream of the kmR start codon and codons 1 to 2 of the kmR ORF, and primers kmR-F3 (5′-gaggatcgtttcgcatgattTTCTTCTGAGCGGGACTCTG-3′) and kmR-R4 (5′-TGGTCCATATGAATATCCTCCTTA-3′) amplified a 268-bp fragment including codons 262–264 of the kmR ORF and 256 bp downstream of the kmR stop codon. Both fragments were ligated by overlapping PCR using oligonucleotides kmR-F1 and kmR-R4 for amplification, and the complementary regions between kmR-R2 and kmR-F3 for overlapping. The resulting fragment, containing the kmR deletion allele, was cloned into pCR2.1 (Invitrogen) and subcloned into the EcoRI site of the suicide plasmid pNPTS138-Cm (Addgene, LGC Standards, Teddington, United Kingdom) to generate plasmid pRCI-65. This suicide plasmid was used to delete the kmR gene of Ba::Tn7wbdRKm^R using the allelic exchange by double recombination (Conde-Álvarez et al., 2006 (link)). Deletion of kmR was checked with oligonucleotides kmR-F1 and kmR-R4.
A Ba::Tn7wbdRΔwbkC mutant potentially expressing only the wbdR encoded acetyltransferase was constructed by PCR overlap using genomic DNA of Ba-parental as template. Primers wbkC-F1 (5′-AGGTGGCGACAAACGAATAA-3′) and wbkC-R2 (5′-GCCCATGCCAATCAAGGT-3′) amplified a 393-bp fragment including codons 1–29 of the wbkC ORF (BAB1_0540), as well as 306 bp upstream of the wbkC start codon, and primers wbkC-F3 (5′-accttgattggcatgggcAGATGGTCGGAAGTCCAGATT-3′) and wbkC -R4 (5′-TCTGAACTCGGCTGGATGAC-3′) amplified a 434-bp fragment including codons 212–259 of the wbkC ORF and 287-bp downstream of the wbkC stop codon. Both fragments were ligated by overlapping PCR using oligonucleotides wbkC-F1 and wbkC-R4 for amplification, and the complementary regions between wbkC-R2 and wbkC-F3 for overlapping. The fragment containing the wbkC deletion allele was cloned into pCR2.1 and subcloned into the BamHI and the XbaI sites of the suicide plasmid pJQK (Scupham and Triplett, 1997 (link)). The resulting mutator plasmid pYRI-31 was used to delete the wbkC gene of Ba::Tn7wbdR by allelic exchange (Conde-Álvarez et al., 2006 (link)). The resulting colonies were screened by PCR with primers wbkC-F1 and wbkC-R4, which amplify a fragment of 827 bp in the mutant and a fragment of 1373 bp in the parental strain.
Bme::Tn7wbdRKm^R was obtained using the modified miniTn7 site-specific integration vector technology (see above). To obtain Bme::Tn7wbdR and Bme::Tn7wbdRΔwbkC the suicide plasmids pRCI-65 and pYRI-31 (see above and Supplementary Table S1) were used.

The genomes were screened for the presence of the 29 CHG profiles defined using the hmmsearch command of HMMER (v. 3.3). Around 1,200,000 hits were found in the 160,987 genomes screened. Results were kept with cutoff values of 10^–3 for e-value and 0.95 for accuracy, and if the length of the predicted proteins was at least 80% of the one of the corresponding profile. When the same protein corresponded to two (or more) overlapping profiles, the result with the lowest e-values was kept. The functions of the proteins screened by HMMER were predicted using InterProScan (v. 5.40, Jones et al., 2014 (link))⁠ and proteins with a predicted function incompatible with aminoglycoside modification were filtered out of the dataset based on a keyword list search. On the one hand, sequences that may have an enzymatic function similar to the ones of AMEs were screened with the following keywords: ‘acetyltransferase’, ‘adenylyltransferase’, ‘adenyltransferase’, ‘phosphotransferase’, ‘phosphoryltransferase’, and ‘nucleotidyltransferase’. On the other hand, sequences whose functions might be involved in aminoglycoside resistance were screened with the following keywords: ‘aminoglycoside’, ‘aminoside’, ‘mycin’, ‘micin’, and ‘amikacin’.
After screening and functional filtering, some CHG profiles did not match any output homologs (i.e. ANTf, APHi, AACk, AACl, and AACn) and they were thus excluded: this left only 24 CHGs. The proteins conserved after this functional filtering step were reclustered with SiLiX (with a minimum of 40% identity over 80% overlap) in order to confirm homology. As some sub-clustered CHGs also contained a low number of hits (2 or 3), only clusters that contained at least 10 sequences were kept (<100 sequences excluded). Three CHGs had to be subdivided and the resulting subclusters will be referred to as their initial CHG name, plus another digit (e.g. AACf1): this additional clustering added 3 CHGs to the 24 remaining ones. This resulted in 46,053 protein sequences predicted to code for AMEs, belonging to 27 CHGs, and spread across 38,523 genomes.

Yeast strains, plasmids, and oligonucleotides used in this study are listed in Tables 1 and 2 and Table S1, respectively. Yeast strains were routinely propagated on YPD or selective solid yeast media at 30°C unless mentioned otherwise. Introduction of plasmids or plasmid fragments into yeast used a lithium acetate-based procedure followed by selection of transformants on appropriate selective solid media (7 (link)).
Several yeast strains were created for this study. yWS2762 (ste24::KAN^R) was created by replacing the STE24 open reading frame in LRB938 with the kanamycin resistance gene (KAN^R). pWS405 was used as the source of KAN^R, as it is flanked by untranslated regions of the STE24 locus that serve as sites for homologous recombination. yWS2802 (ste24::KAN^Rrce1::NAT^R) was created similarly by replacing the RCE1 open reading frame in yWS2762 by recombination with the nourseothricin N-acetyltransferase gene (NAT^R). pWS714 was used as the source of NAT^R as it is flanked by untranslated regions of the RCE1 locus that serve as sites for homologous recombination. Candidates recovered after appropriate selection on YPD solid media containing antibiotic (200 μg/mL G418 or 100 μg/mL nourseothricin) were evaluated using PCR to confirm the presence of the desired disruption. yWS3126 (rce1::NAT^R) was created by cross using LRB938 transformed with pRS316 and yWS2981 (MATα version of yWS2802), followed by random sporulation, screening of haploid candidates for mating type and desired genetic markers and gene disruptions, and propagation on 5-FOA solid medium (0.1% [wt/vol] 5-FOA [Research Products International]) to lose the pRS316 (URA3) plasmid.
Multiple plasmids were created for this study. Construction details are provided in Table S2. The sequences of all engineered plasmids were confirmed by Sanger sequencing for the open reading frames and 5′ and 3′ untranslated regions of encoded genes (Eurofins or Genewiz). Standard ligation or yeast homologous recombination methods were used to create plasmids (7 (link)).