The Human Cytomegalovirus Chemokine vCXCL-1 Modulates Normal Dissemination Kinetics of Murine Cytomegalovirus In Vivo

An adequate in vivo analysis of HCMV’s viral chemokine vCXCL-1 has been lacking. Here we generate recombinant MCMVs expressing vCXCL-1 to study vCXCL-1 function in vivo using MCMV as a surrogate. We demonstrate that vCXCL-1 increases MCMV dissemination kinetics for both primary and secondary dissemination. Additionally, we provide evidence, that the murine neutrophil is largely a bystander in the mouse’s response to vCXCL-1. We confirm the hypothesis that vCXCL-1 is a HCMV virulence factor. Infection of severely immunocompromised mice with MCMVs expressing vCXCL-1 was lethal in more than 50% of infected animals, while all animals infected with parental virus survived during a 12-day period. This work provides needed insights into vCXCL-1 function in vivo.

Mouse cytomegalovirus (MCMV) infection of mice is frequently used to study CMV dissemination. MCMV has similar pathogenesis to HCMV, contains many homologues and orthologues to HCMV genes, and disseminates via innate immune cells (20,21). MCMV does not encode vCXCL-1 but rather encodes a C-C chemokine, MCK2, that enhances dissemination (22)(23)(24). We have previously expressed vCXCL-1 from chimpanzee CMV in MCMV and observed no salivary gland dissemination (25,26), potentially due to inappropriate timing and expression levels of the vCXCL-1 insert. Here, we engineered human vCXCL-1 from the HCMV Toledo strain (vCXCL-1 Tol ) with a 2A peptide linked to the MCMV chemokine MCK2 to ensure appropriate timing and expression levels. We employed the MCMV bacmid of the Smith strain (pSM3fr-MCK2-2fl) in which MCK2 aids viral entry in addition to its role in dissemination (24). Because of MCK2's dual role, we chose to fuse vCXCL-1 Tol to MCK2 instead of deleting it from the recombinant virus. We demonstrate that murine neutrophils are capable of harboring, transferring, and initiating MCMV replication and that CXCR2 stimulation is sufficient to alter MCMV dissemination kinetics. Additionally, we show that infections with recombinant MCMVs expressing vCXCL-1 Tol exhibit increased viral dissemination and virulence.

RESULTS
Murine neutrophils harbor, transfer, and initiate MCMV replication. The capacity of neutrophils to impact cytomegalovirus dissemination is an area of contention. In blood of immunosuppressed patients, neutrophils harbor the largest viral burden (27,28) and transfer infectious HCMV ex vivo and in vitro, but they cannot support productive viral replication (29,30). The interaction between murine neutrophils and MCMV has not been evaluated. Here we use a thioglycolate inflammation model (31) to study the relationship between MCMV and murine neutrophils (Fig. 1A). Mice were infected with an MCMV encoding green fluorescent protein (GFP) under the CMV immediate early (IE) promoter (i.e., 4503) (22), and 4 h postinfection (Fig. 1B), total peritoneal exudate was harvested and analyzed by flow cytometry. Peritoneal exudate cells (PECs) in which virus has entered, uncoated, and expressed IE genes were expected to become GFP positive (GFP ϩ ). All live GFP ϩ cells were initially gated and further analyzed for the presence of neutrophil (Ly6G) and myeloid/granulocyte (CD11b) markers (Fig. 1B). Approximately 2% of all PECs were GFP ϩ , with roughly half this population being CD11b ϩ . Further analysis of the CD11b ϩ subpopulation indicated that ϳ55% of these cells were Ly6G ϩ CD11b ϩ neutrophils and ϳ45% were Ly6G Ϫ CD11b ϩ , likely dendritic cells or patrolling (resident) monocytes (32). The GFP ϩ CD11b Ϫ PECs could be susceptible cell types such as epithelial and fibroblastic cells but not T or B cells (33). The gating strategy used to obtain these results is outlined (see Fig. S1B in the supplemental material). To ensure that the GFP signal is from cells that were infected and expressing GFP and not due to passive uptake or endocytosis of GFP ϩ virions, we infected cultured fibroblasts and used the translational inhibitor cyclohex-imide (34) to demonstrate that GFP expression is de novo and not due to passive uptake of GFP contamination in the viral preparation (Fig. S2).
We further evaluated the contribution of neutrophils to MCMV infection. Using anti-Ly6G microbeads, a 97% pure neutrophil population was prepared from peritoneal exudate from thioglycolate-treated, MCMV-infected mice (Fig. S1A). This population included approximately 25% of the total GFP ϩ cells as described above. The remaining population of peritoneal exudate cells (designated Flowthrough) represented ϳ75% of all the GFP ϩ cells. When both populations were assayed for viral genome via quantitative PCR (qPCR) (Fig. 1C), we did not observe a significant difference in viral genome content between the two populations, indicating that neutrophils and other cell populations harbor MCMV genome 4 h postinfection (p.i.). As neutrophils are phagocytes and may be in the process of destroying virions, we carried out an infectious center assay. Consistent with this notion, the Ly6G ϩ population had dramatically less infectious MCMV compared with Flowthrough (Fig. 1D). To determine whether this neutrophil population was capable of infecting a new host, isolated neutrophils were adoptively transferred into uninfected immunocompetent mice. Three days after neutrophil transfer, blood was harvested and assayed for MCMV genome via qPCR. Transfer Live GFP-expressing cells were gated for CD11b ϩ cells followed by Ly6-G ϩ gating. (C) Total DNA was extracted from 1 ϫ 10 6 purified neutrophils or Flowthrough cells. qPCR was performed to evaluate the number of MCMV genomes present. (D) Infectious center assay for determining the number of infected neutrophils. A total of 1 ϫ 10 6 purified neutrophils were incubated with an uninfected fibroblast monolayer for 5 days, and the number of plaques was evaluated. (E) MCMV genomes were quantified from whole blood isolated from mice that were adoptively transferred 1 ϫ 10 6 neutrophils from infected mice. Values are average titers Ϯ standard deviations (SD) (error bars). Statistical significance was determined by Student's t test and indicated as follows: ns, not significant; ***, P Յ 0.001. Viral Chemokines Enhance MCMV Dissemination ® of infected neutrophils was sufficient to infect a new host (Fig. 1E), although it is possible that a contaminating cell population could contribute this infectivity as well.
CXCR2 stimulation alters normal MCMV dissemination kinetics. Because vCXCL-1 Tol functions primarily through CXCR2 (14,15,35), we sought to understand whether a targeted CXCR2 response could alter normal MCMV dissemination kinetics. When mice are injected intraperitoneally (i.p.) with interleukin 17A (IL-17A), mesothelial cells release host CXCL-1 (i.e., Gro␣) eliciting a neutrophilic, CXCR2-mediated response (36). We confirmed reports that IL-17A treatment alters only the neutrophil population in the peritoneum 4 h following treatment (data not shown). We exploited this experimental setup to mimic the viral chemokine's function in vivo and evaluate the contribution of the vCXCL-1/CXCR2 signaling axis on viral dissemination ( Fig. 2A). Mice treated with IL-17A prior to i.p. infection with MCMV exhibited significantly greater viral burden in spleen and lungs 5 days postinfection (dpi) compared to control animals treated with vehicle control (Fig. 2B and C). IL-17A-treated animals also had greater levels of salivary gland (SG)-associated virus seeding 5 dpi (Fig. 2D). Although there was faster seeding, IL-17A-treated animals did not exhibit a significant difference in viral titers within this organ 14 dpi compared to control animals (data not shown).
Neutrophils were depleted 2 days prior to IL-17A treatment and MCMV infection, then treated every other day after infection with anti-Ly6G or vehicle control (Fig. S3A). The titers of the virus in the spleens were determined 5 dpi (Fig. 2E). Neutrophil depletion resulted in a decreased viral burden in the spleens of IL17A-treated mice, suggesting that in this inflammatory milieu, neutrophils play a role in dissemination perhaps in addition to normal lymphatic drainage from the peritoneum (37).
Generation of recombinant MCMV expressing vCXCL-1 Tol . We have previously generated recombinant MCMVs that overexpress vCXCL-1 or host CXCL-1. These viruses disseminated normally to the spleen, liver, and lung but were defective in SG dissemination (25), leading to a hypothesis that levels and/or timing of chemokine expression hindered SG dissemination. Therefore, we engineered a recombinant MCMV in which vCXCL-1 from the HCMV Toledo strain (vCXCL-1 Tol ) was expressed as a cleavable fusion protein with the MCMV chemokine, MCK2. A picornavirus 2A self-cleaving peptide enabled the coexpression of the two chemokines (38) (Fig. 3A). The recombinant MCMV was generated by coupling classical bacterial recombineering with bacterial artificial chromosome from the Smith strain (pSM3fr-MCK2-2fl) (39) and CRISPR/Cas9 technology. HindIII RFLP analysis confirmed galK insertion into pSM3fr-MCK2-2fl, resulting in a 1.5-kb size increase of the 7.1-kb band into the top of the triplet band (Fig. 3B, boxed in red). The 500-bp addition of vCXCL-1 Tol into the 7.1-kb band resulted in the creation of a doublet in the middle band of the HindIII triplet (Fig. 3B). The supernatants of Viral Chemokines Enhance MCMV Dissemination ® pSM3fr-MCK2-2fl-infected MEF 10.1 cells were subjected to Western blot analysis using anti-FLAG antibodies to identify MCK2 and anti-HIS antibodies to identify vCXCL-1 Tol (Fig. 3C). Glycosylated MCK2 was detected at ϳ40 kDa (40,41) and vCXCL1 was detected at ϳ17 kDa (26). These data indicate the self-cleaving 2A-peptide efficiently generated two independent proteins. In order to evaluate whether insertion of vCXCL-1 Tol alters replication in vitro, both single-step and multistep growth curves were conducted ( Fig. 3D and E). These analyses demonstrate that insertion of the vCXCL-1 Tol gene does not alter viral replication in vitro.

vCXCL-1 Tol expression alters normal MCMV dissemination kinetics.
To test the hypothesis that vCXCL-1 Tol alters normal dissemination, we infected BALB/c mice with pSM3fr-MCK2-2fl expressing vCXCL-1 Tol (called RMvCXCL-1 Tol ) compared to parental pSM3fr-MCK2-2fl (Smith) and a rescued virus (called RMvCXCL-1 Tol RQ). RQ has had the MCK2-vCXCL1 Tol fusion replaced with the pSM3fr-MCK2-2fl mck2 locus and serves as a control for adventitious mutations that could have arisen during recombineering. In addition, whole-genome sequencing was performed on recombinant MCMVs without revealing any unexpected mutations (data not shown). Even though mice inoculated in the footpad (FP) with recombinant MCMVs did not exhibit any difference in viral burden at the site of inoculation 3 dpi (Fig. 4A), RMvCXCL-1 Tol levels were ϳ100-fold higher in the popliteal lymph node at this time compared to either pSM3fr-MCK2-2fl or RMvCXCL-1 Tol RQ (Fig. 4B). Therefore, the addition of vCXCL-1 Tol to MCMV does not alter replication at the infection site, but it does increase the levels of virus in the draining lymph node (LN). Importantly, vCXCL-1 Tol expression dramatically increased secondary dissemination to SG. Plaque assays of whole-organ homogenates revealed ϳ100-fold increased viral levels in SG at 14 dpi compared to either wild-type (WT) or RQ viruses (Fig. 4C). These data indicate that vCXCL-1 Tol significantly increases both primary and secondary dissemination kinetics of MCK2-repaired pSM3fr bacmidderived virus.
It has been reported that different routes of inoculation produce different dissemination patterns (32). To evaluate this possibility with our recombinant viruses, mice were inoculated intranasally (i.n.) (Fig. 4D) or i.p. (Fig. 4E) with RMvCXCL-1 Tol or Smith, and SGs were harvested for plaque assays at 14 dpi. Regardless of the inoculation route, expression of vCXCL-1 Tol enhances MCMV SG dissemination and/or replication.
Analysis of cellular subsets responding to vCXCL-1 Tol in vivo early during infection. Because vCXCL-1 Tol significantly increased dissemination independent of the inoculation site, we evaluated leukocytes recruited to the inoculation site, draining LN, and bloodstream following a FP infection early during infection to identify inflammatory cells potentially contributing to increased dissemination. Mice were infected as in Fig. 4, and FP, LN, and blood samples from infected animals were analyzed by flow cytometry at 3 dpi. In the FP (Fig. 5A), expression of vCXCL-1 Tol induced a significantly higher neutrophil influx compared to control virus, without altering the levels of either inflammatory monocytes (Ly6G Ϫ Ly6C hi CD11b ϩ CD11c Ϫ ) or patrolling monocytes (Ly6G Ϫ Ly6C lo CD11b ϩ CD11c ϩ ) (32). Within the draining LN, all three leukocyte populations were significantly increased by vCXCL-1 Tol compared to control (Fig. 5B), while blood showed no difference (Fig. 5C). There was no difference in total leukocyte influx into the LN and blood (Fig. S4). These data suggest that vCXCL-1 recruitment of neutrophils to the infection site and draining LN or monocytes in LN could be responsible for increased dissemination to SG.

Removal of specific cellular subsets reveals insights into vCXCL-1 Tol function.
Using depleting antibodies, the importance of different cell types was tested. Neutrophil depletion with anti-Ly6G antibody prior to and during MCMV infection ( Fig. 6A and Fig. S3B) reduced the level of vCXCL-1-induced swelling to levels observed with control virus in the absence of depletion (Fig. 6B). Neutrophil depletion also decreased to the level of swelling in controls. Thus, the increase in swelling due to vCXCL-1 Tol was dependent on neutrophils. To examine whether neutrophil depletion alters viral burden in different tissues, LNs were harvested from MCMV-infected mice that had either been neutrophil depleted or left untreated, and virus titers in the draining LN were determined 3 dpi (Fig. 6C). Neutrophil-depleted mice infected with RMvCXCL-1 Tol had less virus in the LN than non-neutrophil-depleted mice, but viral titers remained higher than the viral titers in the controls. These data indicate that neutrophils are partially responsible for vCXCL-1 Tol -mediated dissemination, but they are not the major cell type responding to vCXCL-1 Tol .

vCXCL-1 Tol increases virulence.
Because there is no effective method for depleting monocytes without depleting other cellular subsets (42), we infected NODϪSCIDϪIL-2 gamma chain-deficient (NSG) mice. These mice lack T, B, and NK cells and have defective macrophages and dendritic cells (43). Although these mice are highly immunodeficient, monocytes and neutrophils remain functional. The number of mice surviving daily was recorded and plotted as a Kaplan-Meier survival curve (Fig. 7A). Interestingly, many mice infected with RMvCXCL-1 Tol died on or before 10 dpi, while all mice infected with the control virus survived to 12 dpi at which point SGs and spleens were harvested. These data indicate that vCXCL-1 Tol is a virulence factor, as 58% of mice infected with recombinant MCMVs expressing vCXCL-1 Tol died compared to mice infected with the Smith strain, which all survived. We also observed significantly higher viral titers in the spleens of mice during infection with RMvCXCL-1 Tol compared to the control group (Fig. 7B), even though there was no difference in SG viral titer for mice in both these groups at 12 dpi (Fig. 7C). These data point to either viral load in the spleen, inflammatory infiltrates into the liver (Fig. S5), or another unknown effect of vCXCL-1 Tol as a determinant for mortality in NSG mice. Viral Chemokines Enhance MCMV Dissemination ®

DISCUSSION
As most HCMV infections are asymptomatic, studying its dissemination from primary infection to the establishment of latency in humans has been limited to in vitro and ex vivo analysis (1, 2). These studies revealed that innate immune cells are major reservoirs of HCMV in blood and contribute to viral dissemination (28). HCMV's reliance on innate immune cells has resulted in the evolution of immunomodulatory proteins such as chemokines that are capable of activating and recruiting innate immune cells (13). The discovery of the HCMV chemokine vCXCL-1 led to the hypothesis that vCXCL-1 recruits innate immune cells to the infection site and that these recruited cells contribute to HCMV dissemination (29,44,45). An adequate in vivo analysis of vCXCL-1 function is lacking. Here we employed a mouse model using recombinant MCMV expressing HCMV vCXCL-1 in vivo. A major limitation of the MCMV model and the reason recombinant MCMVs were generated is that some of the immunomodulatory proteins of MCMV and  Fig. 4. (A to C) Footpads (A), lymph nodes (B), and peripheral blood (C) were harvested 3 dpi and evaluated by flow cytometry. Cells were characterized as follows: neutrophils, Ly6G ϩ CD11b ϩ ; inflammatory monocytes (iMonocytes), Ly6G Ϫ Ly6C hi CD11b ϩ CD11c Ϫ ; patrolling monocytes (pMonocytes), Ly6G Ϫ Ly6C lo CD11b ϩ CD11c ϩ . Values are average titers Ϯ SD (error bars). Statistical significance was determined by Student's t test and indicated as follows: ns, not significant; *, P Յ 0.05; **, P Յ 0.01. HCMV differ. MCMV encodes a C-C chemokine which also aids viral entry in MCMV Smith strain (24). However, HCMV encodes C-C and C-X-C chemokines (13), enhancing recruitment of multiple leukocyte subtypes (i.e., monocytes, neutrophils, and NK cells) in vitro (15,16,46). vCXCL-1 attracts neutrophils and potentially NK cells (15,16). These observations suggest that neutrophils could be major players in HCMV dissemination.
To establish our model, we needed to understand MCMV-neutrophil interactions and how/whether neutrophils have a role in MCMV dissemination. This is important because neutrophils transfer HCMV (29), and vCXCL-1 alters human neutrophil functions (15). We show in a thioglycolate-induced model that murine neutrophils harbor and transfer MCMV to new cells although not as efficiently as other cells in the Flowthrough. This could be due to the neutrophil's shorter half-life ex vivo or to the fact that neutrophils have phagocytosed viral genomes that are not infectious. We also show that viral transcription/translation begins in neutrophils (i.e., GFP expression) (Fig. 1), demonstrating that murine neutrophils can harbor MCMV.
Because vCXCL-1 is expressed late in HCMV's life cycle (1), it is likely that infectious viral particles are being released simultaneously with chemokine expression. Therefore, we sought to mimic this environment. IL-17A injected i.p. into mice induces CXCL-1 (i.e., Gro␣) expression (36). CXCL-1 is a host chemokine that uses the CXCR2 receptor (47), inducing primarily a neutrophilic influx into the peritoneal cavity. Stimulating the CXCR2 signaling axis during infection increased viral burden in primary dissemination organs (i.e., spleen and lungs) and resulted in increased SG seeding. However, IL-17A treatment prior to infection did not lead to increased viral burden in SGs 14 dpi. This potentially could be due to a variety of factors, including localization or longevity of CXCR2 stimulation, as IL-17A treatment is transient. Therefore, it is not surprising that IL-17A treatment impacted only primary dissemination. Interestingly, when neutrophils were depleted from the IL-17A inflammation model, there was only a slight decrease in viral burden compared to nondepleted animals. This points to another CXCR2 ϩ cell or another chemokine/cytokine-responsive cell responding to IL-17A. While neutrophils Previously, we have demonstrated that overexpressing vCXCL-1 in the context of MCMV infection did not change primary dissemination but inhibited secondary dissemination (25). We speculated that the timing and quantity of viral chemokine expression induced an abnormal inflammatory environment, resulting in expedited or premature viral clearance. To alleviate this concern, we generated a recombinant MCMV expressing vCXCL-1 at relatively normal physiological times and levels by linking it to the MCMV protein MCK2. When mice were infected with recombinant vCXCL-1 Tol MCMV, there was a significant difference in viral dissemination kinetics for both primary and secondary dissemination compared to controls. Because different inoculation routes have different dissemination mechanisms, it was important to determine whether the phenotype observed was inoculation route dependent (32,51). We showed that the RMvCXCL-1 Tol dissemination phenotype was still present whether mice were inoculated by the more natural inoculation route (i.e., intranasal [51]) or the more common intraperitoneal route. Because vCXCL-1 Tol alters neutrophil functions (15), murine neutrophils were depleted to dissect this relationship in vivo. As with IL-17A neutrophil depletion, neutrophil depletion coupled with recombinant viral infection significantly reduced viral dissemination but did not return viral burden to WT MCMV levels. These data indicate that there is another cellular target through which vCXCL-1 Tol is functioning.
As monocytes have also been shown to express CXCR2 under certain conditions (48,52), the next logical experiment would be to deplete monocytes. Unfortunately, specific depletion of monocytes is not efficient (42). Instead, we chose to utilize the NSG mouse model. These mice are highly immunocompromised and lack mature B, T, and NK cells (43) and have defective macrophages and dendritic cells, which could participate in MCMV dissemination (32,51,53). However, NSG mice have functional monocytes. Because neutrophils were not the major cell type aiding dissemination (Fig. 6), we hypothesized that if viral dissemination of vCXCL-1 Tol -expressing recombinant MCMV were still increased compared to WT MCMV in NSG mice, it would point toward a monocytic response to vCXCL-1 Tol . Surprisingly, infection of NSG mice with RMvCXCL-1 Tol and not strain Smith resulted in death (Fig. 7). This was not expected, but these results support the claim that vCXCL-1 Tol is a virulence factor (12,15,26,54). Additionally, Fig. 7 shows that primary dissemination (i.e., dissemination to the spleen) is significantly different between the two viruses, but there is no difference in secondary dissemination. This could indicate that primary dissemination and secondary dissemination are separate events mediated by different cell types, similar to our findings with CXCL-1-overexpressing MCMVs (25).
While neutrophils are not the major cell type aiding dissemination, our results provide evidence that murine neutrophils contribute to MCMV dissemination. We also report that vCXCL-1 Tol significantly enhances MCMV dissemination and pathogenesis. However, the hypothesis that neutrophils are the sole responders to vCXCL-1 Tol is incorrect. Other CXCR2-positive cells such as monocytes (48,52) and myeloid-derived suppressor cells (55,56) could be other targets for vCXCL-1. This hypothesis supports previous findings in which monocytes are major drivers of CMV dissemination (1,2,32,57). Interestingly, the deaths of NSG mice infected with vCXCL-1 Tol MCMV provide supporting evidence that this protein is a virulence factor albeit in NSG mice. We have previously shown that vCXCL-1 from different HCMV strains vary in their ability to activate and induce neutrophil recruitment (15), but now we have a system to test their virulence in vivo.

MATERIALS AND METHODS
Plasmids. pCas9 was a gift from Luciano Marraffini (Addgene plasmid 42876) (58). The selectable marker was replaced with kanamycin using Gibson Assembly. The p2A-Tol open reading frame (ORF) was synthesized (Genscript) after adding the P2A sequence (Addgene). This plasmid contains 250 bp of MCK2 that is FLAG tagged followed by the 2A peptide and vCXCL-1 from the HCMV Toledo strain, which is 6ϫ HIS tagged. pcDNA-MCMV IE1 was generated as a qPCR standard.
Cells and mice. All experiments were performed with low-passage (Ͻ20) cells. Mouse embryonic fibroblast 10.1 (MEF 10.1) (59) were cultured in Dulbecco modified Eagle medium (DMEM) (Corning) supplemented with 10% Fetalclone III serum (HyClone, Logan, UT), 1% penicillin-streptomycin, and 1% L-glutamine. BALB/c mice were purchased from Jackson Laboratory (Bar Harbor, ME) and housed under specificpathogen-free conditions. Four-to 5-week-old nonobese diabetic severe combined immunodeficient, IL-2 common ␥ chain null (NSG) mice were purchased from Jackson Laboratory (Bar Harbor, ME). The Institutional Animal Care and Use Committees (IACUCs) at the University of Tennessee and University of Cincinnati approved all animal procedures.
Viruses, BAC mutagenesis, and recombinant virus generation. RM4503 virus was a gift from Ed Mocarski, Emory University (22), and MCMV Smith strain virus was derived from the pSM3fr-MCK2-2fl bacterial artificial chromosome (BAC) (originally from B. Adler) was a gift from Chris Benedict (60). MCMV was produced in vitro using MEF 10.1 cells. All viruses were stored at Ϫ80°C until use. Viral titer was assessed by plaque assay (described below) on MEF 10.1 cells.
BAC mutagenesis was performed on pSM3fr-MCK2-2fl BAC by coupling galK recombineering (39) with CRISPR/Cas9 technology. Briefly, Escherichia coli SW105 containing pSM3fr-MCK2-2fl was induced to express lambda red recombinase, and galK was inserted into the MCK2 locus, resulting in pSM3fr-MCK2-2fl-galK. SW105 cells containing pSM3fr-MCK2-2fl-galK were induced by heat and pCas9 with a gRNA targeting galK along with a PCR product containing the MCK2-2A-vCXCL-1 Tol DNA sequence was transformed into them. Transformants were plated on chloramphenicol and kanamycin. The following day, colonies were streaked onto MacConkey agar base (Difco) containing 1% galactose. Colonies that retained the galK gene were pink, while transformants with the desired recombination were white. SW105 cells containing recombinant BACs were further assessed via HindIII restriction fragment length polymorphism (RFLP) analysis, and the 2A-Tol insert was sequenced via Sanger sequencing after PCR amplification.
vCXCL-1 Tol and control BACs were transfected into MEF 10.1 cells using LT1 transfection reagent (Mirus). BAC origin excision was achieved through serial passage as previously described (61). Confirmation of loss of the BAC origin was assessed in purified viral particles instead of infected cells and then evaluated with PCR. Approximately 10 serial passages were needed to effectively excise the BAC origin.