supporting information - pnas...2011/01/28 · supporting information cunnac et al....

Supporting InformationCunnac et al. 10.1073/pnas.1013031108SI Materials and MethodsConstruction and Usage of the Vectors for Genomic Gene Replace-ment and Complementation. For the construction of the vectorsused in this study, the sequences of DNA fragments amplified byPCR and used for cloning were systematically verified to ensurethe absence of introduced mutations.The ΔhrcQb-hrcU deletion construct pCPP6201 was obtained

by first amplifying the genomic ΔhrcQb-hrcU deletion region fromCUCPB5113 with P1296/P2203 and cloning the resulting PCRproduct digested with BsrBI into the SmaI site of pK18mobsacB.The Ω-SpR cassette was cloned out of this intermediate constructwith an XmnI and EcoRV digest, and the Flp recombinase target(FRT)-GmR cassette from pCPP5209 (GenBank accession no.EU024549) amplified using P2259/P2260 was inserted as a SmaIfragment. pCPP6201 was used to delete hrcQb-hrcU fromCUCPB5585 and create CUCPB5589.pCPP5893 was created by PCR amplification of hopI1 flanking

regions with P2590/P2591 and P2592/P2593 primer pairs. ThePCR fragments were digested with XmaI and ligated with T4 li-gase. The ligation product was gel-purified, digested with EcoRIand NheI, and cloned into EcoRI- and NheI-digested pK18mob-sacB. pCPP5610was used to delete hopI1 and create CUCPB5513.The deletion was confirmed by PCR with P2587/P2588.pCPP5913 was created by PCR amplification of hopB1 flanking

regions with P2679/P2680 and P2681/P2682 primer pairs. ThePCR fragments were digested with XbaI and ligated with T4 li-gase. The ligation product was gel-purified, digested with PstI andEcoRI, and cloned into PstI- and EcoRI-digested pK18mobsacB.pCPP5913 was used to delete hopB1 from CUCPB5560 andcreate CUCPB5565. The deletion was confirmed by PCR withP2685/P2686.pCPP5914 was created by PCR amplification of hopAM1-1

flanking regions with P2609/P2610 and P2611/P2612 primerpairs. The PCR fragments were digested with XbaI and ligatedwith T4 ligase. The ligation product was gel-purified, digestedwith EcoRI and HindIII, and cloned into EcoRI- and HindIII-digested pK18mobsacB. pCPP5914 was used to delete hopAM1-1and create CUCPB5520. The deletion was confirmed by PCRwith P2615/P2616.pCPP5920 was created by PCR amplification of avrPtoB

flanking regions with P2464/P2465 and P2466/P2467 primerpairs. The PCR fragments were digested with EcoRI and ligatedwith T4 ligase. The ligation product was gel-purified, digestedwith BamHI (using a natural recognition sequence present onthe flank) and XbaI, and cloned into BamHI- and XbaI-digestedpK18mobsacB. pCPP5920 was used to delete avrPtoB fromCUCPB5534 and create CUCPB5537. The deletion was con-firmed by PCR with P2677/P2678.pCPP5923 was created in two steps. First, hopAF1 flanking

regions were PCR-amplified with P2468/ P2469 and P2470/ P2471primer pairs. The PCR fragments were digested with XbaI andligated with T4 ligase. The ligation product was gel-purified, di-gested with PstI and BamHI, and cloned into PstI- and BamHI-digested pK18mobsacB. The resulting intermediate construct wassubsequently digested with XbaI, and an SpeI-digested FRT Sp/SmR cassette amplified from pCPP5242 (GenBank accession no.EU024551) with P2257/ P2258 was inserted. pCPP5923 was usedto delete hopAF1 from CUCPB5520. The FRT-flanked antibioticresistance cassettes were removed from the intermediate deletionstrains by transformation and curing of the unstable Flp re-combinase expression vector pCPP5264 to create CUCPB5534.The deletion was confirmed by PCR with P2474/P2475.

pCPP5934 was created by PCR amplification of hopE1 flankingregions with P2479/P2480 and P2481/ P2482 primer pairs. ThePCR fragments were digested with XbaI and ligated with T4 li-gase. The ligation product was gel-purified, digested with SphI,and partially digested with EcoRI, and the full-length productwas cloned into EcoRI- and SphI-digested pK18mobsacB.pCPP5934 was used to delete hopE1 from CUCPB5565 andcreate CUCPB5571. The deletion was confirmed by PCR withP2485/P2486.pCPP5952 was created by PCR amplification of avrPto flanking

regions with P2495/ P2496 and P2497/ P2494 primer pairs. ThePCR fragments were digested with SpeI and ligated with T4 li-gase. The ligation product was gel-purified, digested with EcoRIand PstI (using a natural recognition sequence present on theflank), and cloned into EcoRI- and PstI-digested pK18mobsacB.pCPP5952 was used to delete avrPto from CUCPB5537 andcreate CUCPB5546. The deletion was confirmed by PCR withP2675/P2676.pCPP5953 was created by PCR amplification of a first hopK1

flanking region with P2369/P2368, which was subsequently di-gested with SpeI and XbaI and cloned into the NheI site ofpK18mobsacB to yield an intermediate construct. The secondhopK1 flanking region was amplified with P2367/P2366, digestedwith XbaI and PstI, and cloned into XbaI- and PstI-digested in-termediate construct. pCPP5953 was used to delete hopK1 fromCUCPB5546 and create CUCPB5560. The deletion was con-firmed by PCR with P2619/P2620.pCPP5919 was created by PCR amplification of hopA1 flanking

regions with P2456/ P2457 and P2458/ P2459 primer pairs. ThePCR fragments were digested with XbaI and ligated with T4 li-gase. The ligation product was gel-purified, digested with EcoRI,and cloned into EcoRI-digested pK18mobsacB. pCPP5919 wasused to delete hopA1 from CUCPB5571 and create CUCPB5573.The deletion was confirmed by PCR with P2462/P2463.We initially failed to amplify thePCRflanks for cleandeletionof

hopY1 and therefore, created the hopY1 interruption constructpCPP5983. A FRT Sp/SmR cassette was PCR-amplified frompCPP5242 with P2259/P2260, digested with SmaI, and cloned intoFspI-digested pCPP3417 (pENTR/D/SD::hopY1 ORF). The re-sulting vector was digested with EcoRV and NheI, and the ho-pY1::FRTSp/SmR region was subcloned into SmaI- and XbaI-digested pK18mobsacB to obtain pCPP5983. pCPP5983 was usedto interrupt hopY1 in CUCPB5573 and create DC3000D28E(CUCPB5585). The deletion was confirmed by PCR with P2625/P2626. All 15 previous mutations were reconfirmed to be intact byPCR, and no inversions between FRT sites could be detected.pCPP6214, the native avrPto gene restoration construct, was

built by amplifying an avrPto PCR product encompassing thedeleted region and extending within the recombination flanks ofpCPP5952 deletion construct from DC3000 genomic DNA. Thisfragment was digested with AgeI and XbaI and cloned inpCPP5952 digested with AgeI and XbaI to recreate a WT avrPtolocus with bordering regions for recombination. pCPP6214-mediated restoration in P. syringae strains was systematicallyconfirmed on both sides by colony PCR, with primers pairsoSC461/oSC462 and oSC463/oSC464 designed to anneal on therestored region and a sequence bordering the locus but externalto the neighboring recombination flank.Similarly, pCPP6215, the native avrPtoB gene restoration

construct, was built by cloning a DraIII- and EcoNI-digestedPCR product of the WT avrPtoB locus in the pCPP5920 deletionconstruct digested with DraIII and EcoNI. avrPtoB locus resto-

Cunnac et al. www.pnas.org/cgi/content/short/1013031108 1 of 24

www.pnas.org/cgi/content/short/1013031108

ration was colony PCR-verified with oSC458/oSC457 andoSC460/oSC459.pCPP6216, the conserved effector locus (CEL)/clusterVI ge-

nomic restoration construct was obtained by digesting pCPP3139,which contains the DC3000 genomic region covering the CEL,hrp, and exchangeable effector locus (EEL) clusters, with XbaIand SpeI to release a subgenomic fragment covering the entireCEL and flanking sequences for recombination. The digestionproducts were gel-purified and cloned into XbaI-digestedpK18mobsacB. Identity of the insert was verified by three digestswith different enzymes. CEL/clusterVI genomic restoration wasconfirmed by colony PCR with P0242/P0158 and P1576/P0355.pCPP6217, the shcM-hopM1 genomic restoration construct,

was obtained by cloning a PCR product encompassing the pro-moter and 3′ end sequence of the operon amplified fromDC3000 genomic DNA with oSC473/oSC474 and digested withXbaI in the CEL deletion construct pCPP5734 digested withSpeI. In this vector, the shcM-hopM1 operon is oriented oppositeto the hrpH gene. shcM-hopM1 genomic restoration was con-firmed by colony PCR with P2613/P2501 and P2633/P0242.pCPP6218 was assembled in yeast by homologous re-

combination between a KpnI-digested pK18mobsacB and a poolof PCR products composed of (i) an amplicon encompassing theyeast 2μ origin of replication and the TRP1 selectable markergene from the pYESTrp2 (Invitrogen) plasmid amplified withoSC453 and oSC454 that replaced the sacB gene, (ii) the CYH2counter selectable marker gene amplified from pDEST32 (In-vitrogen) with oSC467 and oSC468, and (iii) the EEL homolo-gous recombination region spanning the PSPTO_1409 codingsequence and amplified from DC3000 genomic DNA with pri-mers oSC469 and oSC470. The CYH2 and PSPTO_1409 regionsin pCPP6218 were sequence-verified. The empty shuttle vectorpCPP6219 was obtained by removing the CYH2 cassette ofpCPP6218 by digestion with XhoI and BamHI followed by T4polymerase filling and self-ligation.

Dual Adapter Recombination and Programmable or Random in VivoAssembly Shuttle System.The first section of this description of theprogrammable or random in vivo assembly shuttle (PRIVAS)system focuses on a presentation of the principles and themethods underlying the system. The second part provides theprotocols that were used for the construction and analysis ofthe artificial gene clusters.

Conceptual and Experimental Design of the System. Overview of theuse of flexible dual adapters, which is central to the PRIVAS system.PRIVAS exploits the ability of short terminal adapters to directrecombination of unrelated DNA fragments in vivo or in vitro. Akey feature of the method is the use of a system of dual adaptersenabling each unique DNA fragment in a set of interest to beflanked by a pair of hybrid universal-flexible adapters. Universaladapters (UAs) are first attached to the DNA fragments by PCR,such that all fragments in the set areflankedononeendbyUA1andthe other by UA2. The flexible adapters (FAs) carry homology toUA1 or UA2, and they also carry unique sequences designed tosupport recombination among themselves and/or with vectorscarrying recombination sites for FA1 and FAn (in a set involvingFA1, FA2, . . . FAn). Because the DNA fragments of interest havebeen universalized with UA1 and UA2 and because FAs froma separately maintained panel of oligonucleotides can be easilyattached by PCR, a small number of starting reagents (the set ofuniversalized DNA fragments and the set of FA oligonucleotides)can be used to generate virtually unlimited random or pro-grammed arrangements of concatenated products using any ofa variety of recombination methods. These recombinationmethods could involve, for example, recombinases functioning invivo or restriction enzymes functioning in vitro (1–4). Dualadapter recombination is used here to support concatenation of

type III effector (T3E) genes through recombination in yeast asa key component of the PRIVAS system.Outline of the PRIVAS system. Dual adapter recombination andPRIVAS were initially conceived as a solution to the need forversatile multigene complementation in various Pseudomonassyringae pv. tomato DC3000 polymutant backgrounds. PRIVASenables the assembly of engineered artificial genetic islandscontaining several (1–5 currently but up to perhaps 10–20) genesor genetic units (GU). The configuration of the islands or clusterscan be randomly or fully specified before construction. For itsimplementation, the system exploits homologous recombinationthrough short (∼35 bp) artificial flexible adapters (FA). Here, weused yeast to perform the recombination reactions for the customassembly of several DNA fragments. The first step of the pro-cedure is PCR amplification of DNA regions of interest usingoligonucleotide primers that are chimeric in that their 3′ endcarries GU-specific homology and their 5′ end carries a short (18–20 bp) UA region. In this work, we amplified a total of 16 GUsfrom the DC3000 genome. The sequences of the gene-specificprimers used are provided in Table S3. Graphical representationsof the corresponding regions in their genomic context are avail-able in Fig. S5.Subsequent secondary PCR reactions used UAFA oligonu-

cleotides that are chimeric in that their 3′ end carries homology toUA1 or UA2 and their 5′ end is rationally designed to yieldFA-flanked GUs that are used as the basic cluster building blockin a homologous recombination reaction.We use the termUAFA(rather than simply FA) here to emphasize that the choice of UA1or UA2 is a design feature in these oligonucleotides. The FAsequences serve as the recombination reaction DNA substratesfor joining separate GUs. In line with our specific goal of in-tegrating these clusters in the genome of P. syringae pv. tomato,the homologous recombination reactions also contain the line-arized shuttle vector pCPP6218 (Fig. S7). The end product ofthese reactions is a circular DNA molecule that includes an ar-tificial gene cluster as well as additional functions allowing rep-lication and selection in yeast andEscherichia coli and selection ofsingle cross-over recombination events at a defined location of theEEL region of a P. syringae pv. tomato DC3000 or derivative cell.Two elementary parameters guide the configuration of artificial

clusters.

The relative orientation and position of the FAs at the extrem-ities of the GUs as encoded by the UAFA oligonucleotidesused in the secondary PCR.

The composition of the mix of DNA fragments (GUs) that arethe substrates of the recombination reaction. Note that if ≥2distinct GUs sharing the same pair of flanking FAs are in-cluded in the assembly reaction, assuming equivalent re-combination efficiencies, the resulting clusters will contain,with equal probability, any of these competing GUs. Thisproperty allows the creation of libraries of random clusters.

As long as the FA flanks of the GUs are designed to ensure thatthere is at least one accessible combination of homologous re-combination events that allows circular closure of the growingrecombinant DNA molecule, all arrangements of GUs and FAunits at any position and in any orientation are theoreticallypossible. To simplify the practical design and implementation ofthe assemblies, a small set of simple conventions has been adopted.

The FA at one extremity of the linearized pCPP6218 vectorhas been designated start (Fig. S7A). The 3′ end strand ofthis extremity is referred to as the forward strand. Conse-quently, a UAFA oligonucleotide whose 5′ sequence is iden-tical to the sequence of this strand at the start FA will bedescribed as forward. This FA is defined as the origin of theassembly, and the corresponding DNA strand is viewed asthe reference strand.


http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1013031108/-/DCSupplemental/pnas.201013031SI.pdf?targetid=nameddest=ST3

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1013031108/-/DCSupplemental/pnas.201013031SI.pdf?targetid=nameddest=SF5




Likewise, the other extremity of the pCPP6218 linearized vec-tor, on the same strand (5′ end of this extremity), has beendesignated end, and the corresponding FA has been namedend; this FA is defined as the end of the assembly.

Besides these external FAs (specified by the vector’s se-quence), other artificial FA sequences were computationallycreated. One strand of each of these sequences has beendesignated as the forward strand.

All assemblies were done such that the forward strand of all ofthe internal FAs was on the same strand as the referencestrand of the vector.

Primary PCR: Generating GUs with UA extensions on both ends. Oligo-nucleotide primers used for primary amplification of geneticelements of interest are composed of a 3′ end specific for thesequence of this element and a 5′ end with the sequence of oneof two invariant UA regions. Fig. 4 summarizes the main featuresof the oligonucleotide pairs and target genetic units as typicallyimplemented in the current version of the PRIVAS system. Onespecification requires that the forward strand of the UA1 ele-ment (AACAGGGAGAGGGTGGTGGT) be appended to thecoding strand of the first downstream coding sequence, about150 bp upstream of predicted promoter regions (hrp box). TheUA2 reverse strand (GGTGGTAGCGGTGCGTAAGT) is ap-pended on the other side, to the 5′ end of the noncoding strand,about 150 bp after the end of the last ORF of the genetic unit (toinclude potential transcription terminator sequences).Secondary PCR: Obtaining GUs flanked with FA homology regions for DNAassembly by recombination. Oligonucleotide primers used for sec-ondary amplification are composed of a UA-specific segment intheir 3′ end and an ∼35-bp homology region in their 5′ end. Thesequences of internal FAs were derived from a computer-generated random 100-kb DNA sequence of 52% GC nucleotidecontent. The FastPCR software (5) by PrimerDigital was used togenerate a list of the best quality (no predicted secondarystructures, no self-annealing, homogenous annealing tempera-ture, and GC content etc.) candidate 35- to 36-bp FA sequencesdrawn from this random molecule. Both UA sequences were, inturn, systematically appended to the 3′ end of this set of candi-date FAs and their reverse complement sequence. Any resultingoligonucleotides that exhibited more than 65% identity with theDC3000 genome were excluded from the set. Again, they weretested for quality, and all possible pairs were inspected for po-tential dimer formation. This provided a sense of how UAFAoligonucleotide sets, composed of combinations of an FA se-quence (forward or reverse strands) and a UA sequence (UA1 orUA2), were likely to perform in PCR reactions with otherUAFA oligonucleotide sets (compatibility). From the in silicoanalysis and selection process of candidate UAFA oligonucleo-tides was derived the FA connection subgraph of Fig. S7B, whichcan be used as a guide for the setup of assemblies of up to 5 GUsin size. An arbitrary strand of each FA has been designated theforward strand and is always located on the forward strand ofthe construct. In addition to assembly paths deemed accessible,the network also depicts the paths that were successfully im-plemented in this version of PRIVAS.In addition to programming the position of GUs inside

a designed cluster, the system also offers the option to flip thegenetic units relative to their bordering FAs by using a pair ofoligonucleotides with the appropriate combinations of UA- andFA-strand sequences. In cases where, for example, the tran-scriptional isolation of individual GUs is to be maximized to avoiddownstream effects, it is possible to assemble GUs in a head tohead or tail to tail pattern so that transcription of two neighboringGUs proceeds in opposite directions. Based on inspection ofnatural effector gene clusters on the DC3000 genome and onpreliminary tests, we systematically programmed our assembliesto achieve this type of configuration with flanking GUs in opposite

orientation and with the GU at position 1 (immediately afterFA_START) in the same orientation as the nptII gene of thevector. Conversely, if one needs to place a given GU under thetranscriptional control of a promoter belonging to another GU,one can create an artificial operon (this is contingent on negli-gible transcription termination activity of the UAFA sequencebridging those GUs, which is likely because these sequences wereselected for minimal secondary structure formation potential).Fig. S7C gives an illustration of how the choice of the UAFA

oligonucleotide pair used in secondary PCR impacts the orien-tation of the targeted GU in subsequent assembly.Tables S3 and S4 list the UAFA oligonucleotide pairs used in

secondary PCR reactions and their respective sequence. It alsoindicates the FA flanks of the resulting GU and their orientationin assemblies.Random assemblies in yeast: Design methodology. We first definea couple of terms that will be used in this section.

Assembly or cluster size is the number of GUs that makes upthe cluster or clusters derived from a specifically designedassembly process.

Bin is a set of GUs sharing the same pair of flanking FAscorresponding to a position within a random cluster.

In the random mode of PRIVAS, the primary parameterdriving the configuration of the resulting clusters is the compo-sition of the pool of GUs transferred into yeast. Hence, the as-sembly design scheme aims at rationally selecting the GUs (andtheir flanking FAs) participating in the recombination reaction soas to achieve specific objectives relative to the properties of theassembled random islands (size, prototypic configuration, degreeof complexity, etc.). For the construction of the cluster librariesscreened in this work, our main goals were to minimize clusterconfiguration biases and maximize the exploration of the avail-able cluster space.In preliminary experiments, we realized that, if DNAmolecules

participating in the in vivo assembly reaction share extensiveidentity outside of the FA sequences, these regions frequentlyundergo homologous recombination as well and cause the per-mutation of flanking FAs. This phenomenon can markedly in-terfere with the specified assembly path, and one of its maineffects is the formation of clusters with a size deviating from theassembly size specified by design. Although this process canfurther increase the diversity of cluster configurations, we con-sidered that this was not a desirable situation, because it decreasescontrol over assembly size and may introduce major biases infavor of a few genetic elements. Therefore, to restrict the po-tential for internal recombination, pools of GUs in random as-semblies were set up so that a given GU was assigned to one andonly one bin.Our random assembly strategy involved multiple independent

parallel assemblies of equal size and different bin compositions,and it attempted to meet the following requirements.

For a given assembly, all available primary PCR products mustbe included, each assigned to a unique bin.

Across the set of parallel assemblies, two primary PCR products should have the same overall probability to fall togetherin a bin or conversely, to appear simultaneously on a randomcluster.

To randomize positional effects, a mechanism ensuring thata given GU is not always located at the same position acrossassemblies should be developed.

Ideally, as many parallel assemblies as necessary should beperformed so that a large fraction of the possible combina-tions of primary PCR products is accessible.

To follow the guidelines stated above, we took advantage ofthe existing algorithmic toolbox of the discipline of design of







experiments. The problem of deciding on bins composition inparallel assemblies can be reformulated in terms of finding a nearoptimal balanced incomplete block design where the number oftreatments v is given the value of the number of distinct primaryPCR products included in the experiment (in this study, v = 15).The number of blocks b is equivalent to the assembly size (i.e.,number of bins, in this study, is 3 or 5); the block size k is aninteger equal to v/k and is equivalent to the constant size of thebins. r, the number of complete replicates of v treatments, isequivalent to the number of parallel assemblies (restricted tofour in this study). The adopted design plans were obtained usingthe Resolvable Row-Columns Designs module (6) of the Gendexsoftware (DesignComputing). This report involves two sets ofassemblies of sizes 3 and 5 GUs. Each set consisted of fourparallel recombination reactions whose compositions followedthe design proposed by the software. After assembly, transferinto E. coli, and conjugation into a P. syringae pv. tomatoDC3000derivative for recombination, colonies on selection media weretransferred in microplates to generate the libraries that weresubsequently partially screened on Nicotiana benthamiana.Inferring random cluster configuration. In cases where the system isused to produce libraries of random clusters for functional assays,the problem of inferring the configuration of clusters of interestinevitably arises. Here, the configuration is taken as an accuratedescription of the identity and orientation of the genetic elementsat each position of the cluster. The composition, however, can bedefined as an unordered list of genetic elements, regardless of theorientation and position. The ideal way to elucidate the clusterconfigurations would be to have access to the error-free completeDNA sequence of the clusters. As this was not feasible in practicefor more than a handful of strains, we conceived a strategy to inferthe cluster configuration of several dozens of strains with ac-ceptable confidence.Considering the primary PCR products that were included in

the construction of the libraries of random clusters of size 3 or 5GUs, it is clear that the theoretical maximal length of a size 5 GUcluster is less than 15 kb, which is within the capabilities of currentcommercial high-performance PCR kits. We, therefore, directlyPCR amplified entire clusters with oligonucleotides annealing onthe conserved external borders of the clusters. The resultingamplicons were subsequently used as templates in several se-quencing reactions primed with oligonucleotides specific for thesequence of the FAs known to be upstream of the variouspositions as specified in the corresponding assembly design. Theexperimental DNA sequence obtained subsequently was used toinfer the identity and orientation of the primary PCR product atthis position in the examined cluster. This procedure is relativelyeasy to perform on dozens or even hundreds of strains, but it islikely that the largest clusters will be amplified with low efficiency,thereby restricting an exhaustive unbiased analysis. Moreover,because only an ∼200- to 900-bp segment of the primary PCRproducts is actually available, downstream rearrangements can-not be formally ruled out. Despite the above limitations, thisprocedure was satisfactorily used to elucidate the hypotheticalconfiguration of the clusters from 56 strains.The inference methodology is based on a simple unambiguous

mapping between (i) the experimental sequences obtained with(ii) a specific FA-specific sequencing oligonucleotide from thecluster PCR product of (iii) a specific strain to (iv) a specificindividual GU. This mapping was obtained by systematicallyquerying a custom BLAST database containing the sequences ofthe amplified primary PCR derived from the DC3000 genomewith the experimental sequences. A few experimental sequencesfailed to produce a hit after this search (Fig. S4), but, in all cases,they aligned with unspecific regions when used to run a BLASTon the DC3000 genome and probably originated from unspecificamplification at the primary PCR stage.

Laboratory Protocols Used in PRIVAS. Primary PCR (amplify and appendUA sequences).Tominimize the introduction ofmutations, the PCRreactions were performed with the high-fidelity PrimeSTAR HSDNA Polymerase from Takara Bio Inc. The primary PCR mixcontained 20 μL 5× PrimStart Buffer, 8 μL provided dNTPs, 2 μLeach GU-specific UA primer synthesized by Integrated DNATechnologies at 10 μM, ∼50 ng of P. syringae pv. tomato DC3000genomic DNA, and 2.5 units of PrimeSTAR HS DNA poly-merase. This 100-μL reaction mix was split into two tubes to carryout independent reactions and decrease the chances that earlymutations predominate after amplification. A typical thermalprogram included an initial denaturation of 1.5 min at 94 °C fol-lowed by a first segment of seven cycles using a touch-down pro-cedure: denaturation at 98 °C for 10 s, annealing at 72 °C, −2 °Cper cycle for 5 s, and extension at 72 °C for 2.5 min. The secondsegment consisted of 23 cycles of 98 °C for 10 s, 56 °C for 5 s, and72 °C for 4 min. The replicate reactions were pooled, and a 30-μLaliquot was run on an agarose gel. The band at the expectedspecific size was purified with the DNA Recovery Kit and Clean-up and Concentrator Kit from Zymo Research. An aliquot wassequenced at the Cornell University Biotechnology ResourceCenter to verify the identity of the amplified DNA fragment.Secondary PCR (append FA sequences). Secondary PCR reactions wereperformed as above except that the second segment involved only16 cycles. A 1/100 dilution of the purified primary PCR fragmentwas used as a template, and the appropriate pair of UAFAprimers, also synthesized by Integrated DNA Technologies, wasincluded in the amplification mix; 3-μL aliquots were systemat-ically run on an agarose gel to verify amplification and specificity,and the rest were stored at −20 °C and used without furtherpurification for yeast transformation and cluster assemblies.Yeast transformation for recombinational assembly of clusters. The yeasttransformation procedure essentially followed the protocolfrom the Yeastmaker transformation system by Clontech. Theyeast strains MaV203 (genotype: MATα; leu2-3,112; trp1-901;his3Δ200; ade2-101; cyh2R; can1R; gal4Δ; gal80Δ; GAL1::lacZ;HIS3UASGAL1::HIS3@LYS2; SPAL10 UASGAL1::URA3) fromInvitrogen were used as a recipient and allowed counter-selection of recircularized pCPP6218 vectors that carry the WTdominant cycloheximide susceptibility allele of the CYH2 gene(7). For a small-scale transformation in a 1.5-mL tube, 75 ngXhoI/SpeI linearized and gel-purified pCPP6218 shuttle vectorand 1.5 μL (50–100 ng) each secondary PCR product were in-cluded in the transformation mix together with the carrier DNA.Directly after heat treatment at 42 °C, the cell pellet was sus-pended in sterile water and plated on synthetic defined (SD)selection media lacking tryptophan with glucose and 5 μg/mLcycloheximide. After 3–4 d at 28 °C, a small-scale transformationproduced more than 2 × 104 colonies on selection plates.Recovery of plasmid DNA from yeast. The OD600 of yeast cells re-suspended from the selection plate or grown overnight in liquidSD media was adjusted to ∼3–4 in 250 μL P1 buffer from theQIAprep Spin Miniprep Kit of Qiagen and 5 mg/mL lyticasefrom Sigma-Aldrich. After 1 h incubation at 37 °C and occa-sional mixing, cells were disrupted through two cycles of in-cubation in liquid nitrogen for 30 s followed by 10 min at 65 °C.After the final heat shock, tubes were allowed to cool down toroom temperature, and buffer P2 (250 μL) was added. The restof the procedure followed the protocol provided in the kit andincluded the endonuclease wash step. Plasmid DNA was elutedin 30 μL water.E. coli S17-1 electroporation. Electrocompetent E. coli S17-1 cellswere transformed with 10 μL DNA preparation extracted fromyeast according to standard protocols.Conjugation of P. syringae pv. tomato DC3000 derivatives for single cross-over insertion of the clusters at the EEL. Bacterial conjugations be-tween the donor E. coli S17-1 cells and recipient DC3000 de-rivatives were performed essentially as described in ref. 8. For




generation of random cluster libraries, E. coli S17-1 cells growingon selection plates were resuspended in liquid LB media, and theOD600 was adjusted to 2.0; 200 μL of this suspension were mixedwith an equal volume of an overnight culture of the recipientstrain. The remainder was stored at −80 °C in 15% glycerol.After 3 d on kanamycin selection plates, more than 5,000 colo-nies were obtained per transformation.Colony PCR amplification of integrated clusters.Amplification of entireclusters integrated at the EEL used the Premix Taq (Ex TaqVersion) PCR kit of Takara Bio with primers GCTGCTCC-ATTCCTTCGAGATGC and GCTTTCTACGTGTTCCGCTT-CCTTTAG annealing outside of the external FAs (start and end).Thermal cycling conditions were as follows: a single step at 94 °C

for 2 min, denaturation for 10 s at 98 °C, annealing at 60 °C for30 s, and extension at 72 °C for 13 min for a total of 35 cycles and20 min final extension at 72 °C.ExoSap clean-up of cluster PCR products for sequencing. Before se-quencing, 5 μL PCR reactions containing the cluster ampliconswere treated with 0.25 μL Exonuclease I (20 U/μL) and 0.5 μLAntartic Phosphatase (5 U/μL) from New England Biolabs at37 °C for 30 min followed by heat inactivation for 15 min at 80 °Cto degrade remaining primers and neutralize unincorporateddNTPs. An appropriate sequencing primer from Table S6 wascombined with the resulting DNA solution and sequenced at theCornell University Biotechnology Resource Center.

1. Raymond CK, Pownder TA, Sexson SL (1999) General method for plasmid constructionusing homologous recombination. Biotechniques 26:134–138.

2. Li MZ, Elledge SJ (2007) Harnessing homologous recombination in vitro to generaterecombinant DNA via SLIC. Nat Methods 4:251–256.

3. Engler C, Gruetzner R, Kandzia R, Marillonnet S (2009) Golden gate shuffling: Aone-pot DNA shuffling method based on type IIs restriction enzymes. PLoS One 4:e5553.

4. Sleight SC, Bartley BA, Lieviant JA, Sauro HM (2010) In-fusion BioBrick assembly and re-engineering. Nucleic Acids Res 38:2624–2636.

5. Kalendar R, Lee D, Schulman AH (2009) FastPCR software for PCR primer and probedesign and repeat search. Genes Genomes and Genomics 3:1–14.

6. Nguyen NK, Williams E (1993) An algorithm for constructing optimal resolvable row-column designs. Aust J Stat 35:363–370.

7. Raymond CK, Sims EH, Olson MV (2002) Linker-mediated recombinational subcloningof large DNA fragments using yeast. Genome Res 12:190–197.

8. Kvitko BH, et al. (2009) Deletions in the repertoire of Pseudomonas syringae pv.tomato DC3000 type III secretion effector genes reveal functional overlap amongeffectors. PLoS Pathog 5:e1000388.




Fig. S1. Overview of the P. syringae pv. tomato DC3000 type III effector (T3E) gene repertoire and key steps and strains used in the disassembly of therepertoire and reassembly of a minimal functional repertoire for virulence in the model plant N. benthamiana. WT DC3000 does not grow well or cause diseasein N. benthamiana. Step 1: Deletion of the hopQ1-1 avirulence determinant enables DC3000 to become virulent on N. benthamiana and all subsequent ex-periments use strains lacking this gene (1). Step 2: Redundant effector groups (REGs) are revealed by the strong reductions in growth observed with deletion ofcertain combinations of T3E genes, such as avrPto/avrPtoB or avrE/hopM1/hopR1, which contrasts with the weaker (if any) reductions attending mutation ofany single effector gene (not depicted) or the deletion of many genes in other combinations (CUCPB5459) (2). Step 3: Deletion of all clusters of well-expressedT3E genes strongly reduces but does not abolish growth relative to the T3SS− mutant (2). Step 4: Deletion/disruption of the remaining well-expressed T3E genesproduces DC3000D28E, which is functionally effectorless but otherwise WT, and grows worse than T3SS− derivatives of either DC3000 or DC3000D28E (Figs. 1and 2). Step 5: Integration into native loci of genes encoding the AvrPto and AvrE REGs reveals AvrPto and AvrPtoB to function as early-acting effectors andpotentiate the virulence contributions of AvrE REG effectors such as HopM1 (CUCPB6016) (Fig. 3). Step 6: Using the programmable or random in vivo assemblyshuttle system (PRIVAS) in the random mode (Fig. 4), 18 T3E genes expressed from native promoters are assembled in groups of three or five in the ex-changeable effector locus (EEL) of CUCPB6016, and multiple combinations are observed to stimulate growth in planta (Fig. S4). Step 7: PRIVAS operating in theprogrammed mode is used to assemble a minimal functional repertoire of T3E genes that enable near WT growth and symptom production in N. benthamiana(CUCPB6032) (Fig. 5).

1. Wei C-F, et al. (2007) A Pseudomonas syringae pv. tomato DC3000 mutant lacking the type III effector HopQ1-1 is able to cause disease in the model plant Nicotiana benthamiana. PlantJ 51:32–46.

2. Kvitko BH, et al. (2009) Deletions in the repertoire of Pseudomonas syringae pv. tomato DC3000 type III secretion effector genes reveal functional overlap among effectors. PLoSPathog 5:e1000388.




Fig. S2. Growth of DC3000D28E and DC3000ΔhopQ1-1 in rich King’s B medium (KB) (1) and Hrp minimal medium (HrpMM) (2). Fresh plates were used to startseed cultures in liquid KB that were then grown to log phase (OD600 ≤ 1.0). The seed cultures were used to initiate 60 mL cultures in KB (A) or HrpMM (B) ata starting OD600 of 0.1. The average OD600 ± SD of triplicate cultures is shown. Growth of DC3000D28E and CUCPB6032 (DC3000D28E and eight effector genes)was also compared with DC3000 and DC3000ΔhopQ1-1 in KB (C). Cultures were prepared and analyzed in the same manner as described above. The ex-periment was repeated three times with similar results.

1. King EO, Ward MK, Raney DE (1954) Two simple media for the demonstration of pyocyanin and fluorescin. J Lab Clin Med 44:301–307.2. Huynh TV, Dahlbeck D, Staskawicz BJ (1989) Bacterial blight of soybean: Regulation of a pathogen gene determining host cultivar specificity. Science 245:1374–1377.



Fig. S3. Validation of the PRIVAS system using hopM1 and avrPtoB. Assays for growth of the indicated strains in N. benthamiana were performed as in Figs. 1,3, and 5. The bars reflect the means and SD of colony-forming unit counts at 6 d postinoculation (dpi) calculated from four replicate leaves. Means with thesame letters are not statistically different (α = 0.05) based on a Tukey’s honestly significant difference (HSD) test. In both experiments, strains were inoculatedby N. benthamiana leaf infiltration of 3 × 104 cfu/mL. The grid below the bar plot describes the genotype of the strains both at the EEL (upper one-half) andnative genomic loci (lower one-half). A gray-filled cell indicates that the corresponding locus is either absent (EEL) or deleted (genome), and a white fillsignifies that the locus is either WT (genome) or integrated at the EEL. A Y in the vector integration row denotes that the pCPP6219 plasmid or derivative wasintegrated at the EEL. (A) Complementation of shcM-hopM1 at the EEL using PRIVAS fully restores the growth of the polymutant strain CUCPB5515 defectivefor the AvrE REG. Population levels of CUCPB6024 carrying the shcM-hopM1 at the EEL are not different from those of CUCPB5440, which harbors the entirenative CEL and are about 1.5 logs higher than the AvrE REG-defective strains CUCPB5515 or CUCPB6023 (CUCPB5515 + pCPP6219). CUCPB5460 serves asa reference for maximal growth in N. benthamiana. (B) Complementation of avrPtoB alone or avrPtoB and shcM-hopM1 at the EEL using PRIVAS phenocopyWT restoration of the corresponding genes at their native genomic location. CUCPB6026 carrying avrPtoB at the EEL in a DC3000D28E background accumulatesto the same extent as its natively restored counterpart CUCPB6012. Likewise, the growth of CUCPB6027 (avrPtoB and shcM-hopM1 at the EEL in DC3000D28E) isindistinguishable from CUCPB6017. CUCPB5460, CUCPB5113, and CUCPB5585 serve as references in N. benthamiana for, respectively, maximal growth, T3SS-independent background growth, and minimal growth exhibited by DC3000D28E.



Fig. S4. Growth in N. benthamiana of CUCPB6016-PRIVAS strains containing GUs randomly assembled from a pool of 15 GUs and identification of effectorgenes underlying different growth phenotypes. (A) A representative example of the growth dat a obtained with one of the four batches of 44 strains from theCUCPB6016-PRIVAS library screened for enhanced growth at 6 dpi. Strains are ordered by increasing mean population levels (n = 2) with error bars representingthe SD. (B) The cluster composition and growth phenotypical classes of a subset of selected strains. After growth profiling as illustrated in A for each of the fourbatches (experiment ID column), ∼10 strains from each tail of the distribution (least and best performing) were selected for PCR amplification of the entireengineered cluster. The table summarizes the composition in GUs and the in planta growth data for 56 of this set of selected strains whose clusters could be

Legend continued on following page



successfully PCR-amplified and the underlying GUs subsequently identified by partial sequencing. Strains that have their ID background colored in yellowpromoted a visible chlorosis in the inoculated leaf area at 10 dpi. Blue (GUs) or light salmon shading indicates that the feature in the corresponding column wasdetected in the cluster. For GUs that contain several effectors, the name of first effector gene of the operon is used in the column heading. Numbers indicatethe number of times that the feature was found on the cluster in distinct locations, and the bottom row shows the total number of occurrences of the featureacross the set of clusters. 1, When sequencing reactions with FA-specific oligos failed to produce an exploitable sequence, the nature of the downstream GUwas deemed not available. 2, Sequencing reactions produced DC3000 sequences outside of our pool of GUs; these illegitimate GUs presumably derive fromnonspecific amplification at the primary PCR stage. a, Partial internal deletion of the HopI1 coding sequence. b, Chimeric hopA1-hopY1. c, Complete HopY1coding sequence deleted, but flanking sequences seem unaffected.



Fig. S5. (Continued)



Fig. S5. Amplified type III effector (T3E) genes of P. syringae pv. tomato DC3000 displayed in genomic context. The figure provides graphical displays forrelevant T3E genes on the chromosome and on plasmid pDC3000A. The following web addresses enable visualization at different scales of these genes on thechromosome and plasmid, respectively, using the National Center for Biotechnology Information genome browser:

http://www.ncbi.nlm.nih.gov/nuccore/28867243?report=graph&m=4880997!,1510850!,3468369!,1507479!,6085532!,1730162!,1115371!,648655!,548241!,5353481!,647504!,5416567!,60025!,82161!,5345062!,4881927!,1513658!,3470377!,1509346!,6087577!,1731665!,1116506!,649879!,550769!,5355333!,648619!,5418354!,61636!,83593!,5348677!&mn=hopE1Start,hopM1Start,hopAB2Start,hopAA1-1Start,hopA1Start,hopAF1Start,hopAM1-1Start,hopC1Start,hopF2Start,hopG1Start,hopH1Start,hopI1Start,hopK1Start,hopY1Start,hopAO1Start,hopE1End,hopM1End,hopAB2End,hopAA1-1End,hopA1End,hopAF1End,hopAM1-1End,hopC1End,hopF2End,hopG1End,hopH1End,hopI1End,hopK1End,hopY1End,hopAO1End&v=38131:101128&c=333333&theme=Default&flip=false&select=null&content=3&color=0&label=0&geneModel=1&decor=0&layout=0&spacing=0&alncolor=on

http://www.ncbi.nlm.nih.gov/nuccore/29171478?report=graph&m=19413!,22550!,14710!,16308!&mn=hopO1-1Start,hopO1-1End,hopX1Start,hopX1End&v=11681:25691&c=3366FF&theme=Default&flip=false&select=null&content=1&color=0&label=0&geneModel=1&decor=0&layout=0&spacing=0&alncolor=on


http://www.ncbi.nlm.nih.gov/nuccore/29171478?report=graph%26m=19413!,22550!,14710!,16308!%26mn=hopO1-1Start,hopO1-1End,hopX1Start,hopX1End%26v=11681:25691%26c=3366FF%26theme=Default%26flip=false%26select=null%26content=1%26color=0%26label=0%26geneModel=1%26decor=0%26layout=0%26spacing=0%26alncolor=on









Fig. S6. Configuration and colony PCR verification of the programmed clusters harboring various combinations of the hopE1, hopAM1-1, or hopG1 T3E genes.(A) Graphical representation of the configuration of the programmed T3E gene clusters integrated in the P. syringae strains of Fig. 5. The colored arrowssymbolize genes. Their orientation indicates the direction of the transcription of the corresponding operons. The invariable PSPTO_1409 arrow corresponds tothe region of the pCPP6218 shuttle vector undergoing homologous recombination for genomic integration. The expected sizes of the colony PCR amplicons ofthe gene clusters are provided on the right. Critical DNA features used in assemblies are color-coded, and the color keys are provided in the box below. (B)Ethidium bromide-stained agarose gel of the gene clusters described above amplified by colony PCR. The size of the relevant bands of the DNA ladder run onthe right side is indicated.



Fig. S7. PRIVAS vector and pathways for genetic unit (GU) assembly and orientation flipping. (A) Map of PRIVAS vector pCPP6218, which carries replicationorigins and selectable markers for bacteria and yeast, sites for recombination in yeast with FA-flanked GUs, and PSPTO1409 sequences enabling recombinationwith the exchangeable effector locus (EEL) of P. syringae pv. tomato DC3000. (B) Assembly pathways in the secondary PCR compatibility space of selected UAFAoligonucleotides. (C) Illustration of the ability of UA1 and UA2 swaps in UAFA oligonucleotides to flip GU orientations in assemblies.



Table

S1.

Strainsan

dplasm

idsusedin

this

study

Designation

Gen

otype

Relev

antfeatures

Source

P.syringae

strains

DC30

00W

TP.

syringae

pv.

tomatostrain

DC30

00RfR,ApR

1CUCPB

5113

ΔhrcQ

B-hrcU::Ω

SpR

T3SS

−,Sp

R2

CUCPB

5440

ΔhopD1-hopR1::FRT

ΔIV

3CUCPB

5459

ΔhopU1-hopF2

ΔhopC1-hopH1::FRTΔhopD1-hopR1::FRTΔhopAA1-2-hopG1::FRTpDC30

00A−

pDC30

00B−

ΔIΔIIΔ

IVΔIXΔX

4

CUCPB

5460

ΔhopQ1-1

ΔQ

3CUCPB

5500

ΔhopU1-hopF2

ΔhopC1-hopH1::FRTΔhopD1-hopR1::FRTΔav

rE-shcN

ΔhopAA1-2-hopG1::FRT

pDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔX

4

CUCPB

5513

ΔhopU1-hopF2


rE-shcN


ΔhopI1

pDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1

Thisstudy

CUCPB

5515

ΔhopD1-hopR1::FRTΔav

rE-shcN

ΔIVΔCEL

4CUCPB

5520

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1pDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-1

Thisstudy

CUCPB

5534

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTpDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-1Δ

AF1

Thisstudy

CUCPB

5537

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBpDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-1Δ

AF1

ΔPtoB

Thisstudy

CUCPB

5546

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto

pDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-

1ΔAF1

ΔPtoBΔPto

Thisstudy

CUCPB

5560

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto

ΔhopK1pDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-

1ΔAF1

ΔPtoBΔPtoΔK1

Thisstudy

CUCPB

5565

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto

ΔhopK1ΔhopB1pDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-

1ΔAF1

ΔPtoBΔPtoΔK1Δ

B1

Thisstudy

CUCPB

5571

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto

ΔhopK1ΔhopB1ΔhopE1

pDC30

00A−

pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-

1ΔAF1

ΔPtoBΔPtoΔK1Δ

B1Δ

E1Th

isstudy

CUCPB

5573

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto


ΔhopA1::FRT

pDC30

00A−pDC30

00B−

ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-

1ΔAF1

ΔPtoBΔPtoΔK1Δ

B1Δ

E1ΔA1

Thisstudy

CUCPB

5585

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto


ΔhopA1::FRT

hopY1::FRTS

pRpDC30

00A−pDC30

00B−

DC30

00D28

E:ΔIΔIIΔ

IVΔCEL

ΔIXΔXΔI1ΔAM1-

1ΔAF1

ΔPtoBΔPtoΔK1Δ

B1Δ

E1ΔA1Y

1::SpR

Thisstudy

CUCPB

5589

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto


ΔhopA1::FRT

hopY1::FRTS

pRΔhrcQ

B-hrcU::F

RTG

mRpDC30

00A−pDC30

00B−

T3SS

−DC30

00D28

E,Sp

R,Gm

RTh

isstudy

CUCPB

6011

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔhopK1ΔhopB1ΔhopE1

ΔhopA1::FRThopY1::FRTS

pR

pDC30

00A−pDC30

00B−

DC30

00D28

E+Pto,Sp

RTh

isstudy

CUCPB

6012

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPto



pR

pDC30

00A−pDC30

00B−

DC30

00D28

E+PtoB,Sp

RTh

isstudy

CUCPB

6013

ΔhopU1-hopF2


rE-shcN


ΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔhopK1ΔhopB1ΔhopE1


pR

pDC30

00A−pDC30

00B−

DC30

00D28

E+Pto+PtoB,Sp

RTh

isstudy



Table

S1.Cont.

Designation

Gen

otype

Relev

antfeatures

Source

CUCPB

6014

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1ΔhopAA1-2-

hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto



pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1,

SpR

Thisstudy

CUCPB

6015

ΔhopU1-hopF2

ΔhopC1-hopH1::FRTΔhopD1-hopR1::FRTΔhopAA1-2-hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto


ΔhopA1::FRThopY1::

FRTS

pRpDC30

00A−pDC30

00B−

DC30

00D28

E+CEL

,Sp

RTh

isstudy

CUCPB

6016

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1ΔhopAA1-2-

hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav


ΔhopA1::FRT

hopY1::FRTS

pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1+

Pto,Sp

RTh

isstudy

CUCPB

6017

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1ΔhopAA1-2-

hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPto


ΔhopA1::FRT

hopY1::FRTS

pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1+

PtoB,Sp

RTh

isstudy

CUCPB

6018

ΔhopU1-hopF2


ΔhopAM1-1ΔhopAF1

::FRTΔav



pR

pDC30

00A−pDC30

00B−

DC30

00D28

E+CEL

+Pto,Sp

RTh

isstudy

CUCPB

6019

ΔhopU1-hopF2


ΔhopAM1-1ΔhopAF1

::FRTΔav

rPto



pR

pDC30

00A−pDC30

00B−

DC30

00D28

E+CEL

+PtoB,Sp

RTh

isstudy

CUCPB

6020

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1ΔhopAA1-2-

hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1


ΔhopA1::FRThopY1::

FRTS

pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1+

Pto+PtoB,Sp

RTh

isstudy

CUCPB

6021

ΔhopU1-hopF2


ΔhopAM1-1ΔhopAF1



pRpDC30

00A−

pDC30

00B−

DC30

00D28

E+CEL

+Pto+PtoB,Sp

RTh

isstudy

CUCPB

6022

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1EE

L::[pCPP

6219

]ΔhopAA1-2-hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav



pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1+

Pto+EE

L(Ø),Sp

R,Km

RTh

isstudy

CUCPB

6023


rE-shcN

EEL::[pCPP

6219

]ΔIVΔCEL

+EE

L(Ø),Sp

R,Km

RTh

isstudy

CUCPB

6024


rE-shcN

EEL::[pCPP

6219

shcM

-hopM1]

ΔIVΔCEL

+EE

L(M1),Sp

R,Km

RTh

isstudy

CUCPB

6025

ΔhopU1-hopF2


rE-shcN

EEL::[pCPP

6219

]ΔhopAA1-2-

hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto



pRpDC30

00A−pDC30

00B−

DC30

00D28

E+EE

L(Ø),Sp

R,Km

RTh

isstudy

CUCPB

6026

ΔhopU1-hopF2


rE-shcN

EEL::[pCPP

6219

avrPtoB]

ΔhopAA1-2-hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto

ΔhopK1ΔhopB1

ΔhopE1


pRpDC30

00A−pDC30

00B−

DC30

00D28

E+EE

L(PtoB),Sp

R,Km

RTh

isstudy

CUCPB

6027

ΔhopU1-hopF2


rE-shcN

EEL::[pCPP

6219

avrPtoBshcM

-hopM1]


ΔhopAM1-1ΔhopAF1

::FRTΔav

rPtoBΔav

rPto

ΔhopK1

ΔhopB1ΔhopE1


pRpDC30

00A−pDC30

00B−

DC30

00D28

E+EE

L(PtoB+M1),Sp

R,Km

RTh

isstudy

CUCPB

6028

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1EE

L::[pCPP

6219

hopE1

]ΔhopAA1-2-hopG1::FRTΔhopI1

ΔhopAM1-1ΔhopAF1

::FRTΔav

rPto

ΔhopK1ΔhopB1

ΔhopE1


pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1+

PtoB+EE

L(E1

),Sp

R,Km

RTh

isstudy

CUCPB

6029

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1EE

L::[pCPP

6219

hopE1

hopAM1-1]


ΔhopAM1-1ΔhopAF1

::FRTΔav

rPto

ΔhopK1

ΔhopB1ΔhopE1


pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1+

PtoB+EE

L(E1

+AM1-1),

SpR,Km

RTh

isstudy



Table

S1.Cont.

Designation

Gen

otype

Relev

antfeatures

Source

CUCPB

6030

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1EE

L::[pCPP

6219

hopE1

hopG1]


ΔhopAM1-1ΔhopAF1

::FRTΔav

rPto

ΔhopK1ΔhopB1

ΔhopE1


pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1+

PtoB+EE

L(E1

+G1),Sp

R,

Km

RTh

isstudy

CUCPB

6031

ΔhopU1-hopF2


rE-shcN

::shcM

-hopM1EE

L::[pCPP

6219

hopE1

hopAM1-1hopG1]


ΔhopAM1-1ΔhopAF1

::FRTΔav

rPto



pRpDC30

00A−pDC30

00B−

DC30

00D28

E+M1+

PtoB+EE

L(E1

+G1+

AM1-

1),Sp

R,Km

RTh

isstudy

CUCPB

6032

ΔhopU1-hopF2

ΔhopC1-hopH1::FRTΔhopD1-hopR1::FRTEE

L::[pCPP

6219

hopAM1-1hopE1

hopG1]


ΔhopAM1-1ΔhopAF1

::FRTΔav

rPto



pRpDC30

00A−pDC30

00B−

DC30

00D28

E+CEL

+PtoB+EE

L(E1

+G1+

AM1-

1),Sp

R,Km

RTh

isstudy

Plasmids

pK18

mobsacB

pMB1mobnptIIsacB

SucS,Km

R5

pCPP

5264

pRK41

5C1FLP

TcR

6pCPP

5702

pUCP2

6::Ω

KmP a

vrpto

avrPto-cya

Gm

RKm

R6

pCPP

5893

pK18

mobsacB

::ΔhopI1

SucS,Km

RTh

isstudy

pCPP

5913

pK18

mobsacB

::ΔhopB1

SucS,Km

RTh

isstudy

pCPP

5914

pK18

mobsacB

::ΔhopAM1-1

SucS,Km

RTh

isstudy

pCPP

5919

pK18

mobsacB

::ΔhopA1

SucS,Km

RTh

isstudy

pCPP

5920

pK18

mobsacB

::Δav

rPtoB

SucS,Km

RTh

isstudy

pCPP

5923

pK18

mobsacB

::ΔhopAF1

::FRTS

pR

SucS,Km

R,Sp

RTh

isstudy

pCPP

5934

pK18

mobsacB

::ΔhopE1

SucS,Km

RTh

isstudy

pCPP

5952

pK18

mobsacB

::Δav

rPto

SucS,Km

RTh

isstudy

pCPP

5953

pK18

mobsacB

::ΔhopK1

SucS,Km

RTh

isstudy

pCPP

5983

pK18

mobsacB

::hopY1::FRTS

pR

SucS,Km

R,Sp

RTh

isstudy

pCPP

6201

pK18

mobsacB

::ΔhrcQ

b-hrcU::F

RTG

mR

SucS,Km

R,Gm

RTh

isstudy

pCPP

6214

pK18

mobsacB

::avrPto

SucS,Km

RTh

isstudy

pCPP

6215

pK18

mobsacB

::avrPtoB

SucS,Km

RTh

isstudy

pCPP

6216

pK18

mobsacB

::CEL

SucS,Km

RTh

isstudy

1.BuellCR,et

al.(200

3)Th

eco

mplete

gen

omesequen

ceoftheArabidopsisan

dtomatopathogen

Pseu

domonas

syringae

pv.

tomatoDC30

00.ProcNatlAcadSciUSA

100:10

181–

1018

6.2.

Bad

elJL,Sh

imizuR,OhH-S,Collm

erA

(200

6)A

Pseu

domonas

syringae

pv.

tomatoav

rE1/hopM1mutantisseve

rely

reducedin

growth

andlesionform

ationin

tomato.MolPlan

tMicrobeInteract

19:99–

111.

3.W

eiC-F,et

al.(200

7)A

Pseu

domonas

syringae

pv.

tomatoDC30

00mutantlackingthetypeIII

effectorHopQ1-1isab

leto

cause

disea

sein

themodel

plantNicotian

aben

tham

iana.

Plan

tJ51

:32–

46.

4.KvitkoBH,et

al.(200

9)Deletionsin

therepertoireofPseu

domonas

syringae

pv.

tomatoDC30

00typeIII

secretioneffectorgen

esreve

alfunctional

ove

rlap

amongeffectors.PL

oSPa

thog5:e1

0003

88.

5.Schäfer

A,et

al.(199

4)Sm

allmobiliza

ble

multi-purpose

cloningve

ctors

derived

from

theEsch

erichia

coliplasm

idspK18

andpK19

:Se

lectionofdefi

ned

deletionsin

thech

romosomeofCoryneb

acterium

glutamicum..Gen

e14

5:69

–73

.6.

KvitkoBH,R

amosAR,M

orello

JE,O

hHS,

Collm

erA(200

7)Iden

tificationofharpinsin

Pseu

domonas

syringae

pv.

tomatoDC30

00,w

hicharefunctionally

simila

rto

HrpK1in

promotingtran

slocationoftypeIII

secretionsystem

effectors.J

Bacteriol

189:80

59–80

72.



Table S2. Primers used in this study

Primer name 5′ to 3′ sequence

oSC453 TTAGGTCTTTTTTTATTGTGCGTAACTAACTTGCCCGAGGCCCTTTCGTCTTCAAGoSC454 CTCGGTACCCATCGGCATTTTCTTTTGCGTTTTTATTTCTGATTATCAACCGGGGTGGoSC457 GGAACAACAGCACACACAGGoSC458 CAGCCAAGAGGGAAATAAGGoSC459 GATACTGGCTCGGGGTCTGoSC460 ACGGCTCTGGATGGTCGoSC461 CCGTTTGTTATTGGGCGoSC462 AGAGCGATTTGTTGCGAoSC463 CAGGCGTATCAATCAACCAGoSC464 CGTTATCTTCGTCACCCGAGoSC467 TGGGAAGGGCGATCGGTGCGGGCCTCTTCGCTCTCGAGACTAGTAAAGCCTTCGAGCGTCCCoSC468 ATGCATTCGGATCCATATGTGCTAACAACCATTTTGGAGATTCoSC469 ATGGTTGTTAGCACATATGGATCCGAATGCATTGCCAACTGATGoSC470 ATTAATGCAGCTGGCACGACAGGTTTCCCGACTACACAGGGATCGAGCAGAACGCoSC473 GCTCTAGAGTTCCTTTTTTTATATGCCCAACCAACGoSC474 GCTCTAGAGTTAAAACAGCATGAAGCATGCCGGAP0158 TGCGGCAGATCAAACCTTP0242 CGAACAACACAGAGGCTTGGP0355 TTCAGCGATGGCAAGATAAP1296 CACCATGATGATTCGTAGCCTAACP1576 AGAAAGCTGGGTATCATCGCAAGTGAAAGTP2203 CACCATGACCGCACCGATCAAAAP2257 ATTAACTAGTGTGTAGGCTGGAGCTGCTTCP2258 ATTAACTAGTCATATGAATATCCTCCTTAP2259 ATTACCCGGGGTGTAGGCTGGAGCTGCTTCP2260 ATTACCCGGGCATATGAATATCCTCCTTAP2366 ATTAACTAGTAAAATTACGGTGCAGGAGCAGGP2367 TAATTCTAGATCAAGCCGAAGACGACAGACP2368 CACCTCTAGATCTATTCCCCGATTGAGCTAP2369 TAATACTAGTGGTACCTGGTCAGATTCAGTGCP2456 CTGCGAATTCGAGCCCAACGP2457 TAATTCTAGAGCTCATCAGCCTGCTCATCAACGGGGP2458 TAATTCTAGAAAATGAAAGCAGCGTTCGGCGTAAGTGP2459 CGGCGAATTCGAGTTCTGGTTTP2464 ATCTCTAGAGTGCGCGGCCAGAGAATATCP2465 GCCTCGAATTCTCACACCTTTCCCTATACACP2466 ATAGAATTCCCGCGCTGACAGCTAAAAGCCCATP2467 TAATATCTAGAGGACAGGCCGGACTCGATCTP2468 TAATGGATCCTCTGGATGCTGGGTATGTP2469 TAATTCTAGACCCCATGACGGTTCTCTCTTTP2470 TAATTCTAGACAATAATTCAATAAAGCGCTP2471 TAATCTGCAGAAAACTCTACCTCTACGP2474 CTAACCAGATGGCTGTATGCATCCP2475 CTGGGCTTCGATAAAGCGATTCP2479 TAATGAATTCCGGAAATTCGCACCTGATCCAGCAGCP2480 TAATTCTAGAATTCATGCTGATTGCACCCCTAP2481 TAATTCTAGAGACTGAATCCTAGGCTCTGTACGAP2482 TAATGCATGCTCGACCACTTCTCGGTCACGGTCATTP2485 GTGTTCTGCGTCATAGCCTTTGTCP2486 CGATCCAGTTCTCCACAGGCACP2494 TCCTAGAATTCCTTGGTCGAGACCGCCAAGGP2495 TAATGAATTCGCAGCGTAGAACGACAATP2496 TAATACTAGTTCCCATTCGTATACCCTCTTTAGTP2497 TAATACTAGTCAATGATGTCAAGCCGTGTGTGGP2501 CAGCGCCACCTACGATGAGTP2587 AACTCACTGAAGCAGCGCCTTGP2588 CAGGACTGGGGCTCTGGTTTCAP2590 TAATGCTAGCCGGGCAACGCATGCCTTCAATCAGAAP2591 TAATCCCGGGATGAGGCTGGTAATAGGGCATGAGTAP2592 CTTCCCCGGGAACTGATATCGCP2593 GCCTGAATTCACGGCACTGAATP2609 TAATGAATTCTACTGGAGAGGTTGCCACTTP2610 TAATTCTAGAGACTAAAAAACTCAAATCAGAGTGC



Table S2. Cont.

Primer name 5′ to 3′ sequence

P2611 TAATTCTAGAGTGCATGTATGCCTCCAGACGTP2612 CGGAAAGCTTCAAGCCTTTCTCTTCCAGP2613 CCCACCAAGCTGGCTGCATCATP2615 GTCAACGGCCAGGAGCCCTATAP2616 CCGCAAGCGTTCAAGGGTCTP2619 GCGCTCTGTCGCACTAAAGGCAP2620 ATCCTCGCGCGGCATTTGAGP2625 GACGGCCCAAAGAGTCGGTGAAP2626 AGATCGGCCCGATGATGCTCP2633 TAATGGTACCCTGAGTGCGGTGCGGAGCAP2675 CCGTTTGTTATTGGGCGCAAP2676 GCGTATCAATCAACCAGGGCP2677 GCTCGAAGTCAGCGTCAATGP2678 CGGTGAAGTCATCCAGCACTP2679 AGCGCTGCAGACTGATATGGACP2680 TAATTCTAGATCTCATGATTGAATCTCP2681 TAATTCTAGAGTCTGAGCGCTTGAACP2682 TAATGAATTCGGCGTACAGCAGGTCGP2685 AAAGGCAGTCGTCGAGCAGAP2686 CATGGCGTGATACAAGCGG



Table

S3.

Featuresan

dco

rrespondingprimersoftheprimaryPC

Rproduct

included

inPR

IVASassemblie

s

GU

label

Gen

Ban

kaccession

number

ofthe

template

molecu

le

Coordinates

of

theam

plifi

edfrag

men

tson

thetemplate

Amplifi

edgen

omic

region

size

(bp)

Sense

primer

sequen

ceAntisense

primer

sequen

ce

hopA1

NC_0

0457

860

8553

260

8757

72,09

0aa

cagggag

agggtggtggtCGGACAGGTC

ATC

GTG

CAG

ggtggtagcg

gtgcg

taag

tCGAGCGGTT

CTG

TTTA

GCCTT

hopAF1

NC_0

0457

817

3016

217

3166

51,54

0aa

cagggag

agggtggtggtTGCAGTA

TGTA

GGCTT

TTT

GGAGACGA

ggtggtagcg

gtgcg

taag

tTCGGGGCGTT

TGCTT

GGGCCTT

hopAM1-1

NC_0

0457

811

1537

111

1650

61,16

0aa

cagggag

agggtggtggtCGATG

GCGGCGTT

TATG

TGGA

ggtggtagcg

gtgcg

taag

tGCGGGCTA

TTGTT

GAAGGTG

AhopAO1

NC_0

0457

853

4506

253

4867

73,66

0aa

cagggag

agggtggtggtG

CCTT

GTG

GCGGGCTT

GGTG

GT

ggtggtagcg

gtgcg

taag

tTCAGACCCTC

CCTA

TACATT

TAC

TTTC

TATC

ChopC1

NC_0

0457

864

8655

6498

791,27

0aa

cagggag

agggtggtggtG

CTT

TGCCGTC

TTGGCCTA

CTG

Aggtggtagcg

gtgcg

taag

tTCCGATC

TCAGGCGATG

CAATC

CT

hopE1

NC_0

0457

848

8099

748

8192

795

0aa

cagggag

agggtggtggtG

CAACCTG

CTT

TCATT

CCGCT

ggtggtagcg

gtgcg

taag

tCGCTC

GGTG

ATG

CTG

CGTT

hopF2

NC_0

0457

854

8241

5507

692,58

0aa

cagggag

agggtggtggtG

CCCCTT

CGTT

ACCTT

CCAGCGT

ggtggtagcg

gtgcg

taag

tGGATGCGTT

TTGGCGGATG

AC

hopG1

NC_0

0457

853

5348

153

5533

31,87

0aa

cagggag

agggtggtggtA

CCGTC

CAGAGCGTC

GGCAA

ggtggtagcg

gtgcg

taag

tACGAGGAGCGGCCAAGCGGGTA

hopH1

NC_0

0457

864

7504

6486

191,13

0aa

cagggag

agggtggtggtCCTC

GCGTT

TTGCGATA

GTG

Aggtggtagcg

gtgcg

taag

tCGGCGTT

TGTC

TTAATT

CCTT

ChopI1

NC_0

0457

854

1656

754

1835

41,80

0aa

cagggag

agggtggtggtA

GGCTG

AAGATT

TGTG

ACG

CAGAG

ggtggtagcg

gtgcg

taag

tACGCATT

TTTC

CGAGGCAGTG

GA

hopK1

NC_0

0457

860

025

6163

61,63

0aa

cagggag

agggtggtggtCGCATA

AGTG

GCAATC

GGT

ggtggtagcg

gtgcg

taag

tTCAATC

GTA

CCTG

CCTG

TGG

hopM1

NC_0

0457

815

1085

015

1365

82,80

8aa

cagggag

agggtggtggtA

GTT

CCTT

TTTT

TATA

TGCCCAA

CCAACG

ggtggtagcg

gtgcg

taag

tTAAAACAGCATG

AAGCATG

CCGGA

hopO1-1

AE0

1685

519

413

2255

03,20

0aa

cagggag

agggtggtggtG

GAAGGCGACAACATG

CAGAG

ggtggtagcg

gtgcg

taag

tTGCGGATT

GATA

GGTA

TTTT

CACT

hopX1

AE0

1685

514

710

1630

81,60

0aa

cagggag

agggtggtggtG

GGGTC

GCCTC

AGAAAACGGA

ggtggtagcg

gtgcg

taag

tAGCCAAGGCCAAGGGCGTG

AhopY1

NC_0

0457

882

161

8359

31,47

0aa

cagggag

agggtggtggtG

CCAATG

CGTT

TCTC

GATC

Tggtggtagcg

gtgcg

taag

tGCGCTG

CTG

ATG

GGTA

TCTT

avrPtoB

NC_0

0457

834

6836

934

7037

72,10

0aa

cagggag

agggtggtggtCCGTA

TTCTT

ATG

GAAGGGCA

ggtggtagcg

gtgcg

taag

tCAGGTG

CGAAGTC

CGTG

A



Table S4. UAFA oligonucleotide pairs used in secondary PCR reactions

Flanking forward-strand FAs Primer name FA FA strand UA

FA START+FA001 oSC491 FA START F UA1FoSC484 FA001 R UA2R

FA START-FA001 oSC492 FA START F UA2RoSC483 FA001 R UA1F

FA START+FA END oSC490 FA END R UA2RoSC491 FA START F UA1F

FA START-FA END oSC492 FA START F UA2RoSC489 FA END R UA1F

FA001+FA048 oSC534 FA048 R UA2RoSC481 FA001 F UA1F

FA001-FA048 oSC533 FA048 R UA1FoSC482 FA001 F UA2R


FA001-FA002 oSC482 FA001 F UA2RoSC487 FA002 R UA1F

FA001+FA END oSC481 FA001 F UA1FoSC490 FA END R UA2R

FA001-FA END oSC482 FA001 F UA2RoSC489 FA END R UA1F


FA048-FA091 oSC537 FA091 R UA1FoSC532 FA048 F UA2R


FA091-FA002 oSC536 FA091 F UA2RoSC487 FA002 R UA1F

FA002+FA END oSC490 FA END R UA2RoSC485 FA002 F UA1F

FA002-FA END oSC486 FA002 F UA2RoSC489 FA END R UA1F

Dash indicates that GU is flipped.

Table S5. Sequence of UAFA oligonucleotide pairs used in secondary PCR reactions

Primer name Sequence

oSC481 TACGATGCCAGGATTGTGCGATCTTCACGCTCAGGaacagggagagggtggtggtoSC482 TACGATGCCAGGATTGTGCGATCTTCACGCTCAGGggtggtagcggtgcgtaagtoSC483 CCTGAGCGTGAAGATCGCACAATCCTGGCATCGTAaacagggagagggtggtggtoSC484 CCTGAGCGTGAAGATCGCACAATCCTGGCATCGTAggtggtagcggtgcgtaagtoSC485 ACATCTGGCTCACGATATGCCAAACTGCCTCGCCTaacagggagagggtggtggtoSC486 ACATCTGGCTCACGATATGCCAAACTGCCTCGCCTggtggtagcggtgcgtaagtoSC487 AGGCGAGGCAGTTTGGCATATCGTGAGCCAGATGTaacagggagagggtggtggtoSC488 AGGCGAGGCAGTTTGGCATATCGTGAGCCAGATGTggtggtagcggtgcgtaagtoSC489 TGGGAAGGGCGATCGGTGCGGGCCTCTTCGCTCTCGaacagggagagggtggtggtoSC490 TGGGAAGGGCGATCGGTGCGGGCCTCTTCGCTCTCGggtggtagcggtgcgtaagtoSC491 ACGTGTCATCGGTTGCGTCATCGGCTGGGAGCATCaacagggagagggtggtggtoSC492 ACGTGTCATCGGTTGCGTCATCGGCTGGGAGCATCggtggtagcggtgcgtaagtoSC531 GAGTGGACGTTTACAACATCGATCGCCTCGAACCCAaacagggagagggtggtggtoSC532 GAGTGGACGTTTACAACATCGATCGCCTCGAACCCAggtggtagcggtgcgtaagtoSC533 TGGGTTCGAGGCGATCGATGTTGTAAACGTCCACTCaacagggagagggtggtggtoSC534 TGGGTTCGAGGCGATCGATGTTGTAAACGTCCACTCggtggtagcggtgcgtaagtoSC535 GCAGTGTTGGAGTTTTGTACCTCCAGTTGCGGCGAaacagggagagggtggtggtoSC536 GCAGTGTTGGAGTTTTGTACCTCCAGTTGCGGCGAggtggtagcggtgcgtaagtoSC537 TCGCCGCAACTGGAGGTACAAAACTCCAACACTGCaacagggagagggtggtggtoSC538 TCGCCGCAACTGGAGGTACAAAACTCCAACACTGCggtggtagcggtgcgtaagt

FA sequences are capitalized.



Table S6. Oligonucleotide primers used for sequencing

Anneals on reverse strand Primer sequence

FA START ACGTGTCATCGGTTGCGTCFA001 TACGATGCCAGGATTGTGCGFA002 TCACGATATGCCAAACTGCCFA048 GAGTGGACGTTTACAACATCGATCFA091 GCAGTGTTGGAGTTTTGTACCTC



supporting information - pnas...2011/01/28 · supporting information cunnac et al....

Documents