TALE nickase-mediated SP110 knockin endows cattlewith increased resistance to tuberculosisHaibo Wua,b, Yongsheng Wanga,b, Yan Zhanga,b, Mingqi Yanga, Jiaxing Lvb, Jun Liua,b, and Yong Zhanga,b,1

aCollege of Veterinary Medicine and bKey Laboratory of Animal Biotechnology, Ministry of Agriculture, Northwest A&F University, Yangling 712100,Shaanxi, China

Edited by Leif Andersson, Uppsala University, Uppsala, Sweden, and approved February 3, 2015 (received for review November 11, 2014)

Transcription activator-like effector nuclease (TALEN)-mediated ge-nome modification has been applied successfully to create transgenicanimals in various species, such as mouse, pig, and even monkey.However, transgenic cattle with gene knockin have yet to be createdusing TALENs. Here, we report site-specific knockin of the transcrip-tion activator-like effector (TALE) nickase-mediated SP110 nuclearbody protein gene (SP110) via homologous recombination to pro-duce tuberculosis-resistant cattle. In vitro and in vivo challengeand transmission experiments proved that the transgenic cattleare able to control the growth and multiplication of Mycobacteriumbovis, turn on the apoptotic pathway of cell death instead of necrosisafter infection, and efficiently resist the low dose of M. bovis trans-mitted from tuberculous cattle in nature. In this study, we developedTALE nickases to modify the genome of Holstein–Friesian cattle,thereby engineering a heritable genome modification that facil-itates resistance to tuberculosis.

TALEN | homologous recombination | single-strand break | tuberculosis |disease resistance

Gene targeting by homologous recombination can modify thegenome precisely and has been widely used to study gene

function and produce transgenic animals (1–5). Transcriptionactivator-like effector nuclease (TALEN) is a programmable nu-clease that contains a FokI nuclease domain and a DNA-bindingdomain known as “transcription activator-like effector” derivedfrom the plant pathogenic bacteria Xanthomonas spp. TALENinduces a double-strand break (DSB) at a precise, defined po-sition in the genome, resulting in unpredictable gene mutationswhen the DSBs are repaired erroneously by nonhomologous endjoining (NHEJ) (6, 7). However, TALENs also can be used inconjunction with specially designed exogenous donor DNA togenerate large-scale deletions, gene disruptions, DNA additions,or single-nucleotide changes (8, 9). Numerous cases of TALEN-mediated gene knockouts have been reported in the last 2 y (9–12), but successful knockins are rare (7, 13). TALEN-mediatedsite-specific transgenesis has been applied successfully to modelanimals (7, 14–16) and even in large livestock, such as pigs andcattle (17–21). However, to the best of our knowledge, transgeniccattle with gene knockin have yet to be created using TALENs (19).Tuberculosis is a zoonotic disease caused by the transmission

of Mycobacterium bovis from animals to human beings and fromhuman to human (22). It is a serious threat to global publichealth and agriculture (23, 24). Bovine tuberculosis is widelydistributed worldwide, and no effective programs currently existto eliminate or control the disease in many less-developed areasof Africa and Asia (24, 25). Therefore more extensive and ef-fective studies on the control of bovine tuberculosis are urgentlyrequired in these regions. The mouse SP110 gene is emerging asa promising candidate in the control of Mycobacterium tubercu-losis (MTB) infections (26). SP110 can control MTB growth inmacrophages and induce apoptosis in infected cells. In this study,we developed transcription activator-like effector (TALE) nick-ase technology to insert a mouse SP110 gene into the genome ofHolstein–Friesian cattle. Therefore, TALEN represents a vali-dated tool for the targeted genetic modification of this important

livestock species. Moreover, the results of the present studycould contribute to the control of tuberculosis.

ResultsConstruction of TALEN Plasmids and Activity Assessment. In con-sideration of potential synergistic effects of neighboring genes,we designed three active TALENs specific to the intergenic re-gion between surfactant protein A1 (SFTPA1) and methionineadenosyltransferase I alpha (MAT1A) on chromosome 28 (Fig.1A and SI Appendix, Table S1). Because the transfection effi-ciency of TALEN plasmid and mRNA into bovine fetal fibro-blasts (BFFs) is extremely low, EGFP or mCherry was added tothe TALEN vectors with a self-cleaving T2A peptide to sorttransfected cells via flow cytometry (Fig. 1B).The activity of TALENs in human 293-FT cells was screened

with a luciferase single-strand annealing (SSA) assay [pair 1 (9.2 ±0.86) vs. pair 2 (35.8 ± 3.75), P = 0.000; pair 1 (9.2 ± 0.86) vs.pair 3 (39.7 ± 3.17), P = 0.000; pair 2 (35.8 ± 3.75) vs. pair 3 (39.7± 3.17), no significance] (Fig. 1C). The frequency of TALEN-mediated disruption at the target site in BFFs then was determinedby Surveyor nuclease assays (27). Of the three pairs of TALENsdeveloped, pair 2 cleaved the target site most efficiently, asdemonstrated by the increased incidence of allelic mutations(NHEJ frequency) (Fig. 1 D and E). Therefore, pair 2 was usedfor subsequent experiments. To confirm further the presence ofnuclease-induced insertions and deletions at the targeted locus,the targeted locus was PCR amplified from the genomic DNAand transformed into Escherichia coli by TA cloning. We ran-domly picked 865 bacterial colonies for sequencing. Deletions orinsertions were detected in 6.13% of the colonies generated from

Significance

Bovine tuberculosis is a chronic infectious disease that affectsa broad range of mammalian hosts. It is a serious threat toagriculture in many less-developed countries. In this study, weintroduced a mutation to the FokI of the right hand of wild-type transcription activator-like effector nuclease and estab-lished a transcription activator-like effector nickase system thatcreates single-strand breaks in the genome. Then we used thissystem to add the mouse gene SP110 to a specific location in thebovine genome and created transgenic cattle with increased re-sistance to tuberculosis. Our results contribute to the control andprevention of bovine tuberculosis and provide a previously un-identified insight into breeding animals for disease resistance.

Author contributions: H.W. and Yong Zhang designed research; H.W., Y.W., Yan Zhang,M.Y., J. Lv, and J. Liu performed research; H.W. and Yong Zhang analyzed data; and H.W.wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

See Commentary on page 3854.1To whom correspondence should be addressed. Email: zhy1956@263.net.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1421587112/-/DCSupplemental.

E1530–E1539 | PNAS | Published online March 2, 2015 www.pnas.org/cgi/doi/10.1073/pnas.1421587112

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

TALEN-transfected cells (SI Appendix, Table S2; representativesequences are shown in Fig. 1F).

TALE Nickase Restricts Repair to the Homology-Directed Repair Pathway.A targeted single-strand break (SSB) has the potential to restrictrepair to the homology-directed repair (HDR) pathway (28, 29).Zinc-finger nickases (ZFNs) are well established for generatingSSBs; more recently, this strategy even has been reported inTALENs (5, 29–32). Therefore, a mutation at the active site(D450A) that abolishes catalytic activity without affecting pro-tein dimerization or DNA recognition was introduced to FokI ofthe right hand of TALENs (28).An in vitro DNA cleavage assay was performed to assess the

activity of TALENs bearing the D450A mutation. A linear 383-bpPCR fragment containing an off-center target site for specificTALENs was digested with in vitro-synthesized TALENs. Theexpected digestion patterns following strand-specific cleavagewhen resolved under nondenaturing and denaturing conditionsare shown in Fig. 2A. Thus, provision of wild-type TALENsynthesized in vitro resulted in efficient double-strand cleavageof the template DNA, regardless of whether the products wereresolved under nondenaturing or denaturing conditions (Fig. 2B,lanes 1 and 3, >71% cleavage efficiency). In contrast, introduc-tion of D450A to the wild-type TALEN (TALE nickase) elimi-nated double-strand cleavage (Fig. 2B, lane 2). The cleavageproducts also were resolved under denaturing conditions to con-firm the strand-specific nicking activity of TALE nickase. Asexpected, one strand of the double-stranded DNA was cleavedinto two smaller fragments, but the other strand was uncleavedand persisted as the full-length template (Fig. 2B, lane 4). Theseresults demonstrate that the D450A mutation of FokI resulted inthe generation of a potent, strand-specific TALEN.

Then we further confirmed that a TALE nickase-mediatedSSB had the potential to restrict repair to the HDR pathway.Wild-type TALEN or TALE nickase was transfected into BFFs,and genomic DNA was extracted after 72 h. Surveyor nucleaseassays were performed. As shown in Fig. 2C, TALE nickasedramatically decreased the DNA-cleaving ability, but no NHEJevents were detected. The targeted locus then was PCR ampli-fied and transformed into E. coli, and 823 bacterial colonies werepicked and sent for sequencing. Compared with the 6.13% ofdeletions or insertions in the wild-type TALEN group, none was

Fig. 1. Activity assessment of TALENs. (A) Schematicrepresentation of the targeting locus. (B) Schematicrepresentation of the TALEN constructs. The lengthsof constructs were calculated based on TALEN-encodingplasmids with 17 modules. (C) Cleavage activity of eachTALEN was measured by luciferase SSA assay in human293-FT cells. Data are presented as mean ± SD and arederived from three independent experiments. **P <0.01. (D) Frequency of allelic mutation as determined bySurveyor nuclease assays. Different amounts of eachTALEN transfected are shown as indicated. (E) The M-Slocus was PCR amplified, and cleavage of the locus wasmeasured using a Surveyor nuclease assay. The degreeof cleavage was quantified and is shown below eachlane. (F) Some of the representative sequences revealeddistinct TALEN-induced insertions and deletions atthe targeted locus. The binding site of TALEN isunderlined in red. Occurrences of deletions and inser-tions are listed on the right. The lowercase letters rep-resent inserted bases.

Fig. 2. TALE nickase induced an SSB at the M-S locus and eliminated theNHEJ repair pathway. (A) Illustration of the expected digestion patternsfollowing TALE nickase cleavage when resolved under nondenaturing anddenaturing conditions. (B) Actual results of SSB when resolved under non-denaturing and denaturing conditions. (C) Cleavage activity of TALE nickasewas measured by Surveyor nuclease assay. TALE nickase decreased DNA cleavageactivity, but no NHEJ events were detected.

Wu et al. PNAS | Published online March 2, 2015 | E1531

AGRICU

LTURA

LSC

IENCE

SPN

ASPL

US

SEECO

MMEN

TARY

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

found in the colonies generated from TALE nickase-transfectedcells (SI Appendix, Table S2). These results suggest that a tar-geted strand-specific nick could be repaired by HDR and thatsuch nicks do not generate the indels characteristic of the NHEJrepair pathway.

Selection of Transgene. Based on the analysis of the tuberculosis-susceptible strain C3HeB/FeJ and the tuberculosis-resistantstrain C57BL/6J, Kramnik and colleagues (26) mapped a genetic

locus (sst1) with a major effect on tuberculosis susceptibility onmouse chromosome 1 and found that the SP110 (also called“Ipr1”) gene mediates innate immunity to tuberculosis. SP110 isup-regulated in tuberculosis-resistant macrophages after infection,but it is not expressed in tuberculosis-susceptible macrophages.We obtained five different splice variants of SP110 in cattle usingrapid amplification of cDNA ends; however, preliminary experi-ments showed that none of the five bovine SP110 variants wasuseful in restricting the multiplication of M. bovis in macrophages

Fig. 3. Targeted and heritable addition of the SP110 gene using TALE nickase. (A) Schematic representation of the gene-targeting vector. (B) Schematicoverview depicting the targeting strategy for SP110. D450A, FokI bearing a D450A mutation. (C) Schematic overview screening the individual colonies. 5j F, lrF, and 3j R, lr R are primers for regions outside the homologous arms; 5j R and 3j F are primers for the targeting vector region. Southern blot probes are shownas red lines; Hind III digestion is used in Southern blot analysis. (D and E) Southern blot analysis of the nine heterozygous donor cells used for SCNT. (D) A6.8-kb band resulting from targeted inclusion of the SP110 cassette was detected in addition to the 5.9-kb wild-type band when probe 1 was used. (E) Only a6.8-kb targeted band was detected when probe 2 was used.

E1532 | www.pnas.org/cgi/doi/10.1073/pnas.1421587112 Wu et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

(SI Appendix, Fig. S1A). In contrast, the mouse SP110 can controlthe multiplication of M. bovis (SI Appendix, Fig. S1A; 96 h, control0.71 ± 0.09% vs. mouse SP110 0.38 ± 0.05%, P = 0.0021) andinduce apoptosis in macrophages from infected cattle (SI Ap-pendix, Fig. S1B; control 11.2 ± 1.3% vs. mouse SP110 24.5 ±3.1%, P = 0.0012). Therefore, we chose to add an SP110 genederived from mouse to a specific location in the bovine genome.

Addition of the TALE Nickase-Mediated Gene at the Specific Locus.The gene-targeting vector pLoxp-SP110-Neo was constructed asshown in Fig. 3A. We used the bovine endogenous macrophagescavenger receptor 1 (MSR1) promoter to direct mouse SP110expression and express SP110 only in bovine macrophages. TheBFFs used for targeting were obtained from three different fe-male Holstein–Friesian dairy cows. These cattle originally wereimported from Canada (BFF1), Australia (BFF2), and the UnitedStates (BFF3). TALE nickases encoding plasmids were cotrans-fectedwith pLoxp-SP110-Neo to introduce an SSBbetweenMAT1Aand SFTPA1 (the M-S locus) in BFFs (Fig. 3B). Stably transfectedcells were screened by 5′-junction (1.49 kb), 3′-junction (1.67 kb), andlong-range (targeted, 5.98 kb; wild-type, 1.64 kb) PCR to confirmthat stable genetic modification of cells was targeted to the intendedspecific site (Fig. 3C and SI Appendix, Table S3). RepresentativePCR results are shown in SI Appendix, Fig. S2.PCR screening of heterozygous colonies will generate two

bands for a single knockin, namely, a 5.98-kb band characteristicof the insertion of the SP110 gene and a 1.64-kb band from thenormal chromosome [Fig. 3C, long range (lr) primers]. Thus, theheterozygous colonies were selected for Southern blot confir-mation (Fig. 3 D and E; probes 1 and 2). Following confirmationof successful insertion, karyotype analysis of each heterozygouscolony was conducted (a typical and representative karyotype isshown in SI Appendix, Fig. S3). A total of 26 heterozygous col-onies with normal karyotype, compact spindle-like cell mor-phology, and rapid growth were considered suitable for somaticcell nuclear transfer (SCNT).

Nuclear Transfer to Produce SP110 Transgenic Cattle. Nine of thetransgenic cell colonies were used as donor cells to producecloned transgenic cattle. A total of 1,580 reconstructed embryoswere cultured in vitro; of these, 465 were developed into blas-tocysts. No significant difference in the blastocyst formation ratewas observed among the different nuclear donor cell lines (Table1). These transgenic blastocysts were transferred into the oviductsof 147 recipient heifers. A total of 23 calves were born, and 13survived longer than 6 mo (Table 1 and Fig. 4A). As shown byjunction PCR and genome-walking analysis, SP110 was integratedat the expected site in all 13 calves (Fig. 4 B–E and Dataset S1).Southern blot demonstrated that the transgenic cattle were het-erozygous for SP110 knockin (Fig. 4 F and G).We then determined whether MSR1-controlled SP110 ex-

pression is restricted to macrophages only. Macrophages sepa-rated from transgenic cattle were used for Western blot. Themacrophages from all transgenic cattle expressed SP110 cor-rectly (Fig. 5A). No SP110 protein was detected in the skin,

muscle, heart, liver, spleen, lung, kidney, or milk from transgeniccattle (Fig. 5B).We inquired whether insertion of the exogenous gene affects

the expression of endogenous genes nearby. Real-time RT-PCRanalysis detected no significant difference in the relative levels ofexpression of surfactant, pulmonary-associated protein D (SFTPD),mannose-binding lectin 1 (MBL1), SFTPA1, or MAT1A genes intransgenic and control cattle (Fig. 5C; primers are listed in SIAppendix, Table S4).

In Vitro and in Vivo Challenge and Transmission Experiments. To es-timate the ability of SP110 transgenic cattle to respond toM. bovisinfection, macrophages from peripheral blood (SI Appendix, Sup-plementary Materials and Methods and SI Appendix, Fig. S4) werechallenged withM. bovis in vitro. The macrophages from transgeniccattle and control macrophages showed different reactions toM. bovis infection. First, the rate ofM. bovismultiplication was lowerin the macrophages from SP110 transgenic cattle than in thecontrol macrophages (96 h: 149.3 ± 15.6% vs. 268 ± 19.2%, P =0.0027; 120 h: 136.4 ± 23.7% vs. 289 ± 17.3%, P = 0.0013) (Fig.5D). Second, we observed a clear distinction in the mechanism ofmacrophage cell death after infection. The control cattle macro-phages showed characteristic necrosis (24.3%) (Fig. 5E, Center),whereas the transgenic cattle macrophages showed remarkableapoptosis (33.0%) (Fig. 5E, Right).An in vivo challenge experiment also was performed to confirm

the ability of SP110 transgenic cattle to resist M. bovis. Threerandomly selected transgenic cattle and three experimental con-trol cattle (derived from the same cells but without the transgene,breed-, sex-, and age-matched with the transgenic cattle) wereinfected with 5 × 104 cfu of M. bovis by endobronchial instillation.Cattle were killed 16 wk postinfection. The organs susceptible toM. bovis, such as lung, tracheobronchial lymph node, mediastinallymph node, spleen, and liver tissues, were evaluated for lesionsbased on a gross pathology scoring system as previously described(33, 34). As shown in Table 2, although only one of three trans-genic cattle presented with no histopathological lesions, insertingSP110 into cattle significantly reduced the pathology associatedwith M. bovis infection by endobronchial instillation (pathologyscore, 6.5 vs. 32.0) (Table 2). After being examined for grosslesions, the entire organ was homogenized and used for bacterialcfu assay. As shown in Fig. 5F, bacterial loads in the organs oftransgenic cattle after infection were reduced significantly com-pared with those in the control group (spleen, 0.75 ± 0.056 × 105

vs. 0.29 ± 0.062 × 105, P = 0.0005; liver, 0.45 ± 0.030 × 105 vs.0.21 ± 0.071 × 105, P = 0.0015). The reduction was especially no-table in the lung, which is the organ primarily susceptible to virulentM. bovis (1.67 ± 0.26 × 105 vs. 0.33 ± 0.087 × 105, P = 0.000).To estimate further the ability of transgenic cattle to resist

tuberculosis, a transmission experiment was performed. Earlystudies showed that cattle-to-cattle transmission of bovine tu-berculosis occurs at a lower rate in animals living in outdoorconditions than in animals sharing a confined airspace (35, 36).Therefore, we performed the transmission experiment in an in-dependent category 3 biosafety accommodation. First, tuberculin

Table 1. In vivo development of cloned embryos from different transgenic cells

Nuclear donor BFF1 BFF2 BFF3 Total

Cell clone SC178 SC208 SC364 SC504 SC598 SC671 SC821 SC892 SC973 —

Embryos cultured 166 175 182 161 173 204 182 170 167 1,580Blastocysts (%) 58 (34.9) 49 (28.0) 54 (29.7) 42 (26.1) 47 (27.2) 51 (25.0) 61 (33.5) 54 ( 31.8) 49 (29.3) 465 (29.5)Recipients 18 15 15 16 15 17 19 16 16 147Pregnancies (%) 9 (50.0) 7 (46.7) 5 (33.3) 4 (25.0) 2 (13.3) 4 (23.5) 8 (42.1) 6 (37.5) 5 (31.3) 50 (34.0)Calves at birth 5 3 1 3 0 1 3 3 4 23Calves surviving at 6 mo 3 2 0 2 0 1 2 2 1 13

Wu et al. PNAS | Published online March 2, 2015 | E1533

AGRICU

LTURA

LSC

IENCE

SPN

ASPL

US

SEECO

MMEN

TARY

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

skin tests were performed. The comparative changes in skin foldof transgenic cattle were all less than 1 mm (SI Appendix, TableS5), showing that the transgenic cattle being assessed were notinfected with M. bovis. It is generally accepted that cellularresponses characterized by CD4+ T-cell–derived IFN-γ are helpfulin diagnosing people or animals containing MTB (33, 37–40).IFN-γ release assays (IGRAs) were conducted to monitor theIFN-γ release level of control and transgenic cattle that fed withtuberculous cattle. Thus, on stimulation with bovine tuberculinpurified protein derivatives (PPD-B), control cattle developedIFN-γ responses within 6 wk of living with tuberculous cattle, andthe responses increased steadily throughout the postinfectionperiod. In contrast, the PPD-B–specific IFN-γ responses in trans-genic cattle challenged with M. bovis were significantly lower thanthose in the control cattle challenged withM. bovis (9 wk: transgenic0.19 ± 0.08 vs. control 1.32 ± 0.12, P = 0.000; 12 wk: transgenic0.21 ± 0.10 vs. control 1.46 ± 0.15, P = 0.000) (Fig. 5G).Further, a more specific assay, namely an MTB-specific

enzyme-linked immunospot (ELISPOT) assay, was performed after

the transmission experiment to confirm our results. As shown inFig. 5H, the average number of spot-forming cells (SFC) wassignificantly lower in transgenic cattle than in control cattle[early secretory antigenic target-6 (ESAT-6): control 401.1 ±234.1 vs. transgenic 3.56 ± 3.4, P = 0.000; culture filtrate protein-10(CFP-10): control 182.6 ± 137.7 vs. transgenic 4.78 ± 4.2, P =0.000]. Moreover, the number of SFC in transgenic cattle was notsignificantly different from that in negative control cattle (ESAT-6:negative control 1.43 ± 1.37 vs. transgenic 3.56 ± 3.4, P = 0.215;CFP-10: negative control 2.70 ± 2.33 vs. transgenic 4.78 ± 4.2,P = 0.437). After the transmission experiment, all animals incontact with M. bovis were killed for postmortem examination.The lungs and lymph nodes were evaluated for lesions usinga gross pathology scoring system adapted from Vordermeier andWaters, et al. (33, 34). As shown in Table 3, six of nine transgeniccattle presented with no visible or histopathological lesions. Inaddition, a significant reduction in the gross pathology of thelungs and lymph nodes was observed in the transgenic animals(pathology score, 4.7 ± 2.1 vs. 17.8 ± 4.8, P = 0.000) (Table 3).

Fig. 4. Assessment of transgenic cattle. (A) Photographs of SP110 gene-targeted calves that lived longer than 6 mo. The legends in the photographs identifythe origin of donor cell lines. (B–D) 5′-junction (B, 1.49 kb), 3′-junction (C, 1.67 kb), and long-range (D, wild type: 1.64 kb; targeted: 5.98 kb) PCR to confirmsite-specific targeting in transgenic cattle. Templates for PCR were genomic DNA extracted from cattle peripheral blood. Con, control normal cattle. Lanes1–13 represent the 13 live calves. Lanes 14–17 represent four randomly selected dead transgenic calves. (E) Nucleotide sequence between endogenous andexogenous DNA corresponding to homologous recombination in transgenic cattle. (F and G) Southern blot analysis of the genomic DNA extracted fromtransgenic cattle. Lanes 1–13 represent the 13 live calves. Lanes 14–17 represent four randomly selected dead transgenic calves.

E1534 | www.pnas.org/cgi/doi/10.1073/pnas.1421587112 Wu et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

H&E staining also was used to assess the degree of lung pa-thology present. A representative H&E stain of hilar lymph nodeis shown in Fig. 5I. The development of large necrotic lung lesionsafter infection, a characteristic of control cattle, was prevented intransgenic cattle.

The SP110 Transgene Is Heritable and Is Expressed in OffspringMacrophages. A significant concern in the cultivation of trans-genic animals is whether the transgene is maintained in offspring.In this study, we acquired three offspring calves of the transgeniccattle by means of artificial insemination, and one calf was con-

firmed to be heterozygous for SP110 knockin by Southern blot(Fig. 6 A and B). Western blot was performed to examine whetherthe expression of SP110 was maintained in the heterozygousoffspring. The results indicated that SP110 was expressed in themacrophages of the offspring animal, and there was no signifi-cant difference in the level of SP110 expression in the offspringand founder cattle (Fig. 6C). Furthermore, an in vitro challengeexperiment was performed to estimate the ability of macro-phages from the heterozygous offspring to resist tuberculosis. Asshown in Fig. 6D, the rate of M. bovis multiplication was lower inthe macrophages from the heterozygous offspring than in the

Fig. 5. Assessment of the ability of transgenic cattle to resist tuberculosis. (A) SP110 was expressed correctly in macrophages isolated from transgenic cattle.Lanes 1–13 represent the 13 live transgenic cattle. Con, control cattle. (B) SP110 was expressed only in macrophages. Organs were obtained from a pool ofdead transgenic cattle. Milk was obtained from three live transgenic cattle. Con, control donor cells; MP, macrophages. (C) The addition of SP110 did notaffect the expression of nearby endogenous genes. Macrophages were separated from transgenic cattle (n = 9) or control cattle (n = 9). The relative ex-pression levels of SFTPD, MBL1, SFTPA1, and MAT1A were detected by real-time RT-PCR. Each sample was tested individually, but data were analyzed bygroup. (D) Multiplication of M. bovis in macrophages from control (n = 9) or transgenic (n = 9) cattle in vitro. The macrophages were separated from eachanimal individually and were mixed by group. M. bovis multiplication was determined by cfu assays. (E) Flow cytometry analysis of the mechanism of celldeath of the transgenic cattle macrophages infected with M. bovis. Early apoptotic [annexin V+ propidium iodide (PI)−] late apoptotic (annexin V+ PI+), andnecrotic (annexin V− PI+). (Left) Normal macrophages. (Middle) Infected experiment control macrophages. (Right) Infected transgenic macrophages. (F)M. bovisbacterial loads in the organs of the transgenic cattle after endobronchial infection. (G) Amount of IFN-γ produced in experimental control (n = 9) andtransgenic (n = 9) cattle that shared a confined airspace with positive control cattle for 12 wk. (H) Concentrations of ESAT-6 and CFP-10 IFN-γ–producing SFCsin PBMCs of control and transgenic cattle. (I) H&E stains show a tubercle in the hilar lymph node of the control cattle (A and C) and normal tissue of transgeniccattle (B and D) 16 wk after infection. Arrows show the Langhans giant cells in the tubercle. (Magnification: 100× in I, a and b; 400× in I, c and d.) (Scale bars:50 μm.) The transgenic cattle were divided into three groups according to their origin (derived from three different BFFs), and three cattle were pickedrandomly from each group for the experiments presented in C, D, G, and H. Data are shown as mean ± SD and are derived from at least three independentexperiments. NC, negative control; PC, positive control. *P < 0.05; **P < 0.01.

Wu et al. PNAS | Published online March 2, 2015 | E1535

AGRICU

LTURA

LSC

IENCE

SPN

ASPL

US

SEECO

MMEN

TARY

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

control macrophages (96 h: 160.3 ± 14.6% vs. 275.2 ± 21.7%,P = 0.0047; 120 h: 148.5 ± 18.3% vs. 295 ± 17.4%, P = 0.0020).The distinction in the mechanism of macrophage cell death afterinfection was observed also. The control macrophages showednecrosis (control 23.8 ± 3.4% vs. heterozygous 3.3 ± 1.5%, P =0.000), whereas the macrophages from the heterozygous offspringshowed characteristic apoptosis (control 9.1 ± 1.5% vs. hetero-zygous 32.5 ± 4.1%, P = 0.000) (Fig. 6E) after infection withM. bovis. These data indicated that SP110 site-specific knockin fortuberculosis resistance in cattle is heritable.

DiscussionAs of this writing, ZFN, TALEN, and RNA-guided engineerednuclease (RGEN) are the three most widely used and mostpromising tools for genome modification. However, each nucleasehas its own advantages and disadvantages. ZFNs, the first pro-grammable nuclease for genome modification, have been usedand improved in academia and industry for the last two decades(41, 42). For example, clinical trials on the ZFNs for CCR5, themost common coreceptor for HIV-1, have been underway forseveral years, and the therapeutic benefits are very promising

(43, 44). However, despite continuous improvements in ZFNtechnology, a substantial proportion of ZFNs fail, whether theyare produced by design or selection (45–47). In contrast,TALENs can be designed to target almost any given DNA sequenceand achieve a very high success rate. Because a single mismatchbetween modules and base pairs can decrease binding signifi-cantly, TALENs are generally less toxic and more specific thanZFNs. RGEN derives from an adaptive immune system that iswidespread among bacteria and archaea. The unique advantagesof RGENs over ZFNs and TALENs are their simplicity and thefact that they are readily multiplexed; however, specificity remainsan issue in this system (48–52). TALENs and RGENs possiblymay replace ZFNs for routine research based on principles ofsimplicity, efficiency, and reliability. However, we still need toprove whether these nucleases will have similar or greater utilitythan ZFNs. More importantly, in addition to the features of thesethree classes of nucleases, cell type and delivery method also havegreat effects on the activity and success rate. No reliable rulescurrently exist to predict nuclease activity before experimentalvalidation.

Table 2. Gross pathology of transgenic cattle challenged with M. bovis by endobronchialinstillation

AnimalNo. of lobesinfected*

Lungscore

No. of lymphnodes infected†

Lymph nodescore

Total pathologyscore Mean‡

Transgenic 1 2 4 3 4 8 6.5Transgenic 2 1 2 2 3 5Transgenic 3 0 0 0 0 0Control 1 5 21 6 14 35 32.0Control 2 4 15 8 18 33Control 3 4 14 6 14 28

*Lung lobes (left apical, left cardiac, left diaphragmatic, right apical, right cardiac, right diaphragmatic, andright accessory lobes) were examined for lesions using a gross pathology scoring system.†Lymph nodes (mandibular, parotid, medial retropharyngeal, mediastinal, tracheobronchial, hepatic, mesen-teric, and prescapular lymph nodes) were examined for lesions using a gross pathology scoring system.‡Median values per group (n = 3). Only animals with lesions were taken into account.

Table 3. Gross pathology of transgenic cattle challenged by transmission experiment

AnimalNo. of lobesinfected*

Lungscore

No. of lymphnodes infected†

Lymph nodescore

Totalscore Mean ± SD‡

Transgenic group 1 1 2 1 1 3 4.7 ± 2.10 0 0 0 00 0 0 0 0

Transgenic group 2 3 3 2 4 72 2 1 2 40 0 0 0 0

Transgenic group 3 0 0 0 0 00 0 0 0 00 0 0 0 0

Control group 1 4 12 6 13 25 17.8 ± 4.84 13 5 10 233 10 3 8 18

Control group 2 4 12 4 10 223 9 4 9 182 6 3 6 12

Control group 3 3 8 4 9 173 7 2 5 122 7 3 6 13

*Lung lobes (left apical, left cardiac, left diaphragmatic, right apical, right cardiac, right diaphragmatic, and right accessory lobes)were examined for lesions using a gross pathology scoring system.†Lymph nodes (mandibular, parotid, medial retropharyngeal, mediastinal, tracheobronchial, hepatic, mesenteric, and prescapularlymph nodes) were examined for lesions using a gross pathology scoring system.‡Median values per group (n = 9). Only animals with lesions were taken into account.

E1536 | www.pnas.org/cgi/doi/10.1073/pnas.1421587112 Wu et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

TALEN technology has been applied successfully to createtransgenic animals in most small model animal species (10, 11,15). Carlson et al. (17) reported successful TALEN-mediatedgene knockout in bovine embryos. Huang et al. (53) and Lillicoet al. (54) created TALEN-mediated transgenic pigs. Most re-cently, Proudfoot et al. (21) reported that they have createdmyostatin gene-knockout sheep and cattle using TALENs. Herewe report, for the first time to our knowledge, TALE nickase-mediated gene insertion via homologous recombination to pro-duce transgenic cattle. We introduced a D450A mutation to theFokI of the right hand of wild-type TALEN and established aTALE nickase system, which primarily creates a SSB in the ge-nome. A targeted SSB has the potential to restrict repair to theHDR pathway, thereby eliminating the NHEJ pathway and greatlyimproving the efficiency of targeting.The M-S locus was selected for gene targeting for the following

reasons. First, macrophages express many surface receptors thatfacilitate the binding of microorganisms. SFTPA1, one of the sur-factant proteins, may modulate the activity of one or more receptorsthat are responsible for direct binding to M. bovis (55–57); SFTPD,an important paralog of SFTPA1, interacts with compounds, suchas bacterial lipopolysaccharides, in the immune response (58).MBL1 recognizes mannose and N-acetylglucosamine in manymicroorganisms, and it can activate the classical complementpathway (59). Given these facts, we hypothesized that SP110, inconjunction with endogenous genes nearby, may activate anti-M. bovis mechanisms in macrophages (this hypothesis needs tobe explored further experimentally). Second, the chromatin en-coding these genes is activated in macrophages because of theimportant functional role of this gene cluster. Therefore, in-sertion of SP110 in this region could avoid exogenous gene si-lencing caused by chromatin inactivation.We created 13 transgenic cattle using M-S locus-targeted het-

erozygous colonies as donor cells. Heterozygous colonies of cellswith SP110 knockin to a single chromosome 28 will retain onenormal chromosome, which will be helpful for the survival oftransgenic animals. This strategy has been proved in our pre-

viously published work (5), mainly because the normal chromo-some would counteract the defects or neutralize the side effectsintroduced by genome modification.The cleavage efficiency of TALENs at the M-S locus was

lower than previously reported (5, 17). This region may beheavily methylated or even inactivated in BFFs because of itsfunction. TALEN cleavage occurs largely during the S phase ofthe cell cycle when all genomic sequences are exposed for rep-lication. To confirm our hypothesis, the FSCN1-ACTB (F-A)locus was selected. Higher cleavage efficiency was achieved, butrelatively lower blastocyst rates, pregnancy rates, and birth rateswere observed also (SI Appendix, Supplementary Result, Fig. S5,and Tables S6 and S7). These data suggested that the F-A locusprobably is not a safe harbor for the transgene. Although wedetected an off-target mutation in one of 19 F-A locus gene-targeted cell clones (SI Appendix, Fig. S6 and Tables S8 and S9),so far there is no convincing evidence that the potential toxicity isassociated with off-target effects. The underlying mechanismsshould be examined further in future investigations.Based on in vitro and in vivo challenge and on the trans-

mission experiments, the SP110 transgenic cattle could controlthe growth and multiplication of M. bovis, activate the apoptoticpathway of cell death instead of necrosis after infection withM. bovis, and efficiently resist the low dose of M. bovis transmittedfrom tuberculous cattle in nature. In this study, we also acquiredthree offspring calves of the founder transgenic cattle, and one calfwas heterozygous for SP110 knockin. We found that SP110 isexpressed in the heterozygous offspring, and an in vitro challengeexperiment proved that tuberculosis resistance is maintained inthe macrophages from the heterozygous calf (Fig. 6). All theseresults demonstrate that inserting SP110 into cattle is a highlypromising technique for creating resistance to M. bovis infectionand that this genome modification for tuberculosis resistance incattle is heritable.M. bovis can evade host immune defense by inducing necrosis

rather than by inhibiting the apoptosis of macrophages (60). Inthe present study, SP110 transgenic cattle could activate the

Fig. 6. The SP110 transgene is heritable and is expressed in offspring macrophages. (A and B) Southern blot analysis of three offspring cattle using probe 1(A) and probe 2 (B). The results show that one of the offspring is heterozygous for the SP110 knockin. Lanes 1–3 represent the three offspring cattle. NC,negative control; PC, positive control. (C) The expression level of SP110 was detected by Western blots (WB). Lanes 1–3 represent the three offspring cattle.GAPDH serves as a loading control. (D) In vitro multiplication ofM. bovis in the macrophages from control cattle, control offspring, or heterozygous offspring.M. bovis multiplication was determined by cfu assay. Control offspring are the two offspring cattle without the SP110 transgene. (E) Apoptosis and necrosisrates of control and offspring macrophages infected with M. bovis were determined by flow cytometry.

Wu et al. PNAS | Published online March 2, 2015 | E1537

AGRICU

LTURA

LSC

IENCE

SPN

ASPL

US

SEECO

MMEN

TARY

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

apoptotic pathway of cell death instead of necrosis after infectionwith M. bovis. Therefore, previously undescribed mechanisms ofSP110 or SP110-interacting proteins may exist, in which SP110 orSP110-interacting proteins control various aspects of the mac-rophage life activities, including activation, differentiation, andimmune response to pathogens. It would be interesting to in-vestigate further the specific mechanism of SP110 in determiningthe fate of macrophage. Moreover, other factors associated withSP110 should be examined, and their potential roles of tuber-culosis resistance should be determined.

Materials and MethodsEthics Statement. This study was carried out in strict accordance with theguidelines for the care and use of animals of Northwest A&F University. Allanimal experimental procedures were approved by the Animal Care Com-mission of the College of Veterinary Medicine, Northwest A&F University.Bovine ovaries from slaughtered mature cows were collected from theTumen abattoir, a local abattoir in Xi’An, China. Six-month-old tuberculosis-free calves were obtained from Keyuan Cloning Co., Ltd. and were kept inthe Animal Services Unit in a category 3 biosafety accommodation. Everyeffort was made to minimize animal pain, suffering, and distress and toreduce the number of animals used.

Surveyor Nuclease Assay. The capacity of each TALEN for native gene dis-ruption activity at its target locus was determined by Surveyor nuclease(Transgenomic) assay in BFFs. In brief, genomic DNA from TALEN-treated cellswas extracted using a Universal Genomic DNA Extraction Kit (TakaRa). Thetargeted loci were PCR amplified using the following primers: pair 1, F (5′-GAGAAGGAAATGGCAACCCAC-3′) and R (5′-CGGAAATCTTGATTCCAGCTT-3′); pair 2, F (5′-CCTTCCGCCTCTGTAGGTACAGA-3′) and R (5′-GTAGGACA-CAGTGCCGCAAACCC-3′); pair 3, F (5′-CGAATTCACTTTCACTTTCGT-3′) andR (5′-CAGTTCTTCACTTTCTGCCATA-3′).

PCR productswere digestedwith Surveyor nuclease and analyzedby agarosegel electrophoresis according to themanufacturer’s instructions. Quantificationwas based on relative band intensities using Image J software.

Cell Culture and Transfection. Primary BFFs were isolated from 1-mo-old fe-tuses. The tissueswereminced, plated on 60-mmPetri dishes (Corning Costar),and cultured with DMEM/F12 (Gibco, Invitrogen) supplemented with 10%(vol/vol) FBS (HyClone) and 10 ng/mL epidermal growth factor. HEK-293FT cells(ATCC) were cultured with DMEM (Invitrogen) supplemented with 10%(vol/vol) FBS. BFFs were harvested using 0.25% trypsin/EDTA solution(Invitrogen). Cells (1 × 107) were resuspended in Opti-MEM (Gibco), mixed, ifnot otherwise indicated, with 10 μg of linearized donor plasmid and 5 μg ofeach TALEN-encoding plasmid, and electroporated at 510 V with three pulsesof 1-ms duration using the BTX Electro-cell manipulator ECM2001 (BTX Tech-nologies). Electroporated cells were sorted via flow cytometry and plated on10-cm plates at 1 × 106 cells per plate. Individual colonies were selected andexpanded after G418 selection (600 ng/mL) 10–14 d after electroporation.

Nuclear Transfer. Ovaries were collected from the local abattoir and trans-ported to the laboratory within 4–6 h in sterile saline at 20 °C. In vitromaturation of oocytes, enucleation, microinjection, and fusion of recon-structed oocytes were carried out in our laboratory according to the pre-viously described methods (5). The reconstructed oocytes were cultured untilthey developed to blastocyst stage. Three or four fresh day 7 blastocystsproduced in vitro were nonsurgically transferred to randomly assignedsynchronized recipient heifers on day 7 after estrus. Pregnancy was di-agnosed by rectal palpation on day 35 and confirmed by ultrasonography onday 60 after blastocyst transfer.

M. bovis Challenge and Transmission Experiments. For the challenge experi-ment, three control and three transgenic calves were infected with 5 × 104 cfuof M. bovis (strain AF 2212/97) by endobronchial instillation as previously de-scribed (33, 61). At the end of the experimental period, the calves were killedby i.v. injection of sodium phenobarbitone, and postmortems were performed.

We had set a control group and an experimental group in the transmissionexperiment. The control group comprised a negative control (a normal animalwithout the transgene or M. bovis infection) and a positive control (a normalanimal without the transgene but infected with M. bovis by endobronchialinstillation and diagnosed as tuberculous). The experimental group comprisedthe control (cloned animals without the transgene) and transgenic animals.Positive controls used for the transmission experiment were produced by

endobronchial instillation with 5 × 104 cfu of M. bovis. Skin tests and IFN-γassays were performed to confirm that cattle were infected with M. bovis. Thepositive results of bacterial cultures of respiratory fluids proved that infectedcattle could transmit M. bovis into the environment. Positive controls werereconfirmed by the presence of tuberculous lesions in the lungs and lymphnodes through postmortems after the transmission experiment. For thetransmission experiment, nine positive controls, nine experimental controls,and nine transgenic cattle were fed together in a confined accommodationfor 12 wk. Blood samples were collected, and IFN-γ assays were performed tomonitor the level of IFN-γ release at the time points indicated in Fig. 5H.

Postmortem and Pathology Scoring System. Postmortems were performedafter the challenge and transmission experiments. Lung lobes (left apical, leftcardiac, left diaphragmatic, right apical, right cardiac, right diaphragmatic,and right accessory lobes) were examined externally for the occurrence oflesions, followed by slicing of the lung into 0.5- to 1-cm-thick slices that thenwere examined individually for lesions. Lymph nodes (mandibular, parotid,medial retropharyngeal, mediastinal, tracheobronchial, hepatic, mesenteric,and prescapular lymph nodes) were sliced into 1- to 2-mm-thick slices thatwere examined for the presence of visible lesions. Lungs and lymph nodeswere evaluated using a semiquantitative gross pathology scoring systemadapted from Vordermeier andWaters et al. (33, 34). Lung lobes were scoredindividually based on the following scoring system: 0 = no visible lesions; 1 =no external gross lesions, but lesions seen upon slicing; 2 = fewer than fivegross lesions <10 mm in diameter; 3 = more than five gross lesions <10 mmin diameter; 4 =more than one distinct gross lesion >10 mm in diameter; 5 =gross coalescing lesions. The scores of the individual lobes were added tocalculate the lung score. Lymph node pathology was based on the followingscoring system: 0 = no necrosis or visible lesions; 1 = small focus (1–2 mm indiameter); 2 = several small foci or a necrotic area at least 5 × 5 mm; 3 =extensive necrosis. The scores of lung lobes and lymph nodes were added todetermine the total pathology score per animal. All scoring was performedby the same operator to ensure scoring consistency.

Cfu Assay. Infection with M. bovis was performed by the State Key Labora-tory of Veterinary Etiological Biology (Lanzhou, China). In brief, a bacterialsuspension (∼107 bacteria per 106 cells) was added to the medium and in-cubated at 37 °C and 5% (vol/vol) CO2 for 4 h. Cells then were washed ex-tensively with PBS to remove noningested bacteria. At the time pointsindicated in the text after infection, bacterial cfu were quantitated byplating on 7H10 agar plates (Difco Laboratories). Quantitative assessment ofbacterial burden in organs was performed as previously described (34). Inbrief, after being examined for gross lesions, the entire organ was homog-enized in phenol red nutrient broth. The homogenates then were dilutedwith PBS, plated on 7H10 agar plates, and incubated for 8 wk at 37 °C.

IGRAs. IGRAs were performed using a BOVIGAM kit (Prionics AG) according tothe manufacturer’s instructions. In brief, whole-blood samples were in-cubated with PPD-B to stimulate the lymphocytes to secrete IFN-γ. Theplasma supernatants were harvested after 24 h of incubation, and IFN-γ wasestimated using a sandwich enzyme immunoassay. Optical density at 450 nmwas determined using a VICTOR ×5 Multilabel Plate Reader (PerkinElmer).

ELISPOT Assay. MTB-specific ELISPOT assays were performed as previouslydescribed (33, 62). In brief, ELISPOT plates (Millipore) were coated overnightat 4 °C with mouse anti-bovine IFN-γ monoclonal antibody (Thermo Scien-tific). Peripheral blood mononuclear cells (PBMCs) (2 × 105) then were addedand cultured at 37 °C for 24 h. The cells were stimulated with ESAT-6 andCFP-10 peptides in separate wells following procedures performed strictlyaccording to the manufacturer’s recommendations. The response of stimu-lated cultures was considered positive when the test well contained at leastsix more spots than the control well.

Statistical Analysis. Data are presented as the mean ± SD and are derived fromat least three independent experiments. Statistical significances were analyzedusing the Student’s t test. A value of P < 0.05 was considered significant.

ACKNOWLEDGMENTS. We thank Ye Liu, Chengcheng Cui, and Xu Liu forexpert technical assistance; Kun Ru and the ViewSolid Biotech Company forassistance in constructing TALEN plasmids; Xiaoning He for assistance inkaryotype analysis; and the Keyuan Cloning Company for assistance in theM. bovis challenge and transmission experiments. This work was supportedby National Major Project for Production of Transgenic Breeding Grant2013ZX08007-004 and National High Technology Research and DevelopmentProgram of China (863 Program) Grant 2011AA100303.

E1538 | www.pnas.org/cgi/doi/10.1073/pnas.1421587112 Wu et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1

1. Meyer M, de Angelis MH, Wurst W, Kühn R (2010) Gene targeting by homologousrecombination in mouse zygotes mediated by zinc-finger nucleases. Proc Natl Acad SciUSA 107(34):15022–15026.

2. Liu H, et al. (2014) TALEN-mediated gene mutagenesis in rhesus and cynomolgusmonkeys. Cell Stem Cell 14(3):323–328.

3. Sommer D, et al. (2014) Efficient genome engineering by targeted homologous re-combination in mouse embryos using transcription activator-like effector nucleases.Nat Commun 5:3045.

4. Xue H, et al. (2009) A targeted neuroglial reporter line generated by homologousrecombination in human embryonic stem cells. Stem Cells 27(8):1836–1846.

5. Liu X, et al. (2013) Zinc-finger nickase-mediated insertion of the lysostaphin gene intothe beta-casein locus in cloned cows. Nat Commun 4:2565.

6. Mahfouz MM, et al. (2011) De novo-engineered transcription activator-like effector(TALE) hybrid nuclease with novel DNA binding specificity creates double-strandbreaks. Proc Natl Acad Sci USA 108(6):2623–2628.

7. Zu Y, et al. (2013) TALEN-mediated precise genome modification by homologousrecombination in zebrafish. Nat Methods 10(4):329–331.

8. Miller JC, et al. (2011) A TALE nuclease architecture for efficient genome editing. NatBiotechnol 29(2):143–148.

9. Bedell VM, et al. (2012) In vivo genome editing using a high-efficiency TALEN system.Nature 491(7422):114–118.

10. Sung YH, et al. (2013) Knockout mice created by TALEN-mediated gene targeting. NatBiotechnol 31(1):23–24.

11. Gaj T, Gersbach CA, Barbas CF, 3rd (2013) ZFN, TALEN, and CRISPR/Cas-based methodsfor genome engineering. Trends Biotechnol 31(7):397–405.

12. Li T, Liu B, Spalding MH, Weeks DP, Yang B (2012) High-efficiency TALEN-based geneediting produces disease-resistant rice. Nat Biotechnol 30(5):390–392.

13. Hockemeyer D, et al. (2011) Genetic engineering of human pluripotent cells usingTALE nucleases. Nat Biotechnol 29(8):731–734.

14. Katsuyama T, et al. (2013) An efficient strategy for TALEN-mediated genome engi-neering in Drosophila. Nucleic Acids Res 41(17):e163.

15. Huang P, et al. (2011) Heritable gene targeting in zebrafish using customized TALENs.Nat Biotechnol 29(8):699–700.

16. Wang H, et al. (2013) TALEN-mediated editing of the mouse Y chromosome. NatBiotechnol 31(6):530–532.

17. Carlson DF, et al. (2012) Efficient TALEN-mediated gene knockout in livestock. ProcNatl Acad Sci USA 109(43):17382–17387.

18. Tan W, et al. (2013) Efficient nonmeiotic allele introgression in livestock using customendonucleases. Proc Natl Acad Sci USA 110(41):16526–16531.

19. Tan WS, Carlson DF, Walton MW, Fahrenkrug SC, Hackett PB (2012) Precision editingof large animal genomes. Adv Genet 80:37–97.

20. Moghaddassi S, Eyestone W, Bishop CE (2014) TALEN-mediated modification of thebovine genome for large-scale production of human serum albumin. PLoS ONE 9(2):e89631.

21. Proudfoot C, et al. (2015) Genome edited sheep and cattle. Transgenic Res 24(1):147–153.

22. Grange JM (2001) Mycobacterium bovis infection in human beings. Tuberculosis(Edinb) 81(1-2):71–77.

23. Thoen C, Lobue P, de Kantor I (2006) The importance of Mycobacterium bovis asa zoonosis. Vet Microbiol 112(2-4):339–345.

24. Ayele WY, Neill SD, Zinsstag J, Weiss MG, Pavlik I (2004) Bovine tuberculosis: An olddisease but a new threat to Africa. Int J Tuberc Lung Dis 8(8):924–937.

25. Buddle BM, et al. (2003) Revaccination of neonatal calves with Mycobacterium bovisBCG reduces the level of protection against bovine tuberculosis induced by a singlevaccination. Infect Immun 71(11):6411–6419.

26. Pan H, et al. (2005) Ipr1 gene mediates innate immunity to tuberculosis. Nature434(7034):767–772.

27. Guschin DY, et al. (2010) A rapid and general assay for monitoring endogenous genemodification. Methods Mol Biol 649:247–256.

28. Sanders KL, Catto LE, Bellamy SR, Halford SE (2009) Targeting individual subunits ofthe FokI restriction endonuclease to specific DNA strands. Nucleic Acids Res 37(7):2105–2115.

29. Kim E, et al. (2012) Precision genome engineering with programmable DNA-nickingenzymes. Genome Res 22(7):1327–1333.

30. Wu Y, et al. (2014) TALE nickase mediates high efficient targeted transgene integrationat the human multi-copy ribosomal DNA locus. Biochem Biophys Res Commun 446(1):261–266.

31. Ramirez CL, et al. (2012) Engineered zinc finger nickases induce homology-directedrepair with reduced mutagenic effects. Nucleic Acids Res 40(12):5560–5568.

32. Wang J, et al. (2012) Targeted gene addition to a predetermined site in the humangenome using a ZFN-based nicking enzyme. Genome Res 22(7):1316–1326.

33. Vordermeier HM, et al. (2002) Correlation of ESAT-6-specific gamma interferon pro-duction with pathology in cattle following Mycobacterium bovis BCG vaccinationagainst experimental bovine tuberculosis. Infect Immun 70(6):3026–3032.

34. Waters WR, et al. (2009) Efficacy and immunogenicity of Mycobacterium bovisDeltaRD1 against aerosol M. bovis infection in neonatal calves. Vaccine 27(8):1201–1209.

35. Pollock JM, Rodgers JD, Welsh MD, McNair J (2006) Pathogenesis of bovine tuber-culosis: The role of experimental models of infection. Vet Microbiol 112(2-4):141–150.

36. Morris RS, Pfeiffer DU, Jackson R (1994) The epidemiology of Mycobacterium bovisinfections. Vet Microbiol 40(1-2):153–177.

37. Flynn JL, Chan J (2001) Immunology of tuberculosis. Annu Rev Immunol 19:93–129.38. Flynn JL, Ernst JD (2000) Immune responses in tuberculosis. Curr Opin Immunol 12(4):

432–436.39. Delogu G, Li A, Repique C, Collins F, Morris SL (2002) DNA vaccine combinations ex-

pressing either tissue plasminogen activator signal sequence fusion proteins orubiquitin-conjugated antigens induce sustained protective immunity in a mousemodel of pulmonary tuberculosis. Infect Immun 70(1):292–302.

40. Khader SA, et al. (2007) IL-23 and IL-17 in the establishment of protective pulmonaryCD4+ T cell responses after vaccination and during Mycobacterium tuberculosischallenge. Nat Immunol 8(4):369–377.

41. Kim H, Kim JS (2014) A guide to genome engineering with programmable nucleases.Nat Rev Genet 15(5):321–334.

42. Carroll D (2011) Genome engineering with zinc-finger nucleases. Genetics 188(4):773–782.

43. Perez EE, et al. (2008) Establishment of HIV-1 resistance in CD4+ T cells by genomeediting using zinc-finger nucleases. Nat Biotechnol 26(7):808–816.

44. Carroll D (2014) Genome engineering with targetable nucleases. Annu Rev Biochem83:409–439.

45. Ramirez CL, et al. (2008) Unexpected failure rates for modular assembly of en-gineered zinc fingers. Nat Methods 5(5):374–375.

46. Kim JS, Lee HJ, Carroll D (2010) Genome editing with modularly assembled zinc-fingernucleases. Nat Methods 7(2):91–, author reply 91–92.

47. Sander JD, et al. (2010) ZiFiT (Zinc Finger Targeter): An updated zinc finger engi-neering tool. Nucleic Acids Res 38(Web Server issue):W462–468.

48. Mali P, et al. (2013) CAS9 transcriptional activators for target specificity screening andpaired nickases for cooperative genome engineering. Nat Biotechnol 31(9):833–838.

49. Fu Y, et al. (2013) High-frequency off-target mutagenesis induced by CRISPR-Casnucleases in human cells. Nat Biotechnol 31(9):822–826.

50. Cradick TJ, Fine EJ, Antico CJ, Bao G (2013) CRISPR/Cas9 systems targeting β-globin andCCR5 genes have substantial off-target activity. Nucleic Acids Res 41(20):9584–9592.

51. Pattanayak V, et al. (2013) High-throughput profiling of off-target DNA cleavagereveals RNA-programmed Cas9 nuclease specificity. Nat Biotechnol 31(9):839–843.

52. Hsu PD, et al. (2013) DNA targeting specificity of RNA-guided Cas9 nucleases. NatBiotechnol 31(9):827–832.

53. Huang J, et al. (2014) RAG1/2 knockout pigs with severe combined immunodeficiency.J Immunol 193(3):1496–1503.

54. Lillico SG, et al. (2013) Live pigs produced from genome edited zygotes. Sci Rep3:2847.

55. Ernst JD (1998) Macrophage receptors for Mycobacterium tuberculosis. Infect Immun66(4):1277–1281.

56. Pasula R, et al. (1997) Surfactant protein A (SP-A) mediates attachment of Myco-bacterium tuberculosis to murine alveolar macrophages. Am J Respir Cell Mol Biol17(2):209–217.

57. Gaynor CD, McCormack FX, Voelker DR, McGowan SE, Schlesinger LS (1995) Pulmonarysurfactant protein A mediates enhanced phagocytosis of Mycobacterium tuberculosis bya direct interaction with human macrophages. J Immunol 155(11):5343–5351.

58. Leth-Larsen R, et al. (2005) A common polymorphism in the SFTPD gene influencesassembly, function, and concentration of surfactant protein D. J Immunol 174(3):1532–1538.

59. Garred P, Honoré C, Ma YJ, Munthe-Fog L, Hummelshøj T (2009) MBL2, FCN1, FCN2and FCN3-The genes behind the initiation of the lectin pathway of complement. MolImmunol 46(14):2737–2744.

60. Behar SM, Divangahi M, Remold HG (2010) Evasion of innate immunity by Myco-bacterium tuberculosis: Is death an exit strategy? Nat Rev Microbiol 8(9):668–674.

61. Vordermeier HM, et al. (1999) Development of diagnostic reagents to differentiatebetween Mycobacterium bovis BCG vaccination and M. bovis infection in cattle. ClinDiagn Lab Immunol 6(5):675–682.

62. Losi M, et al.; European Tuberculosis Network TBNET (2007) Use of a T-cell interferon-gamma release assay for the diagnosis of tuberculous pleurisy. Eur Respir J 30(6):1173–1179.

Wu et al. PNAS | Published online March 2, 2015 | E1539

AGRICU

LTURA

LSC

IENCE

SPN

ASPL

US

SEECO

MMEN

TARY

Dow

nloa

ded

by g

uest

on

Nov

embe

r 20

, 202

1