Plant Cell Reports

, Volume 35, Issue 7, pp 1451–1468

The expanding footprint of CRISPR/Cas9 in the plant sciences

Review

DOI: 10.1007/s00299-016-1987-x

Cite this article as:
Schaeffer, S.M. & Nakata, P.A. Plant Cell Rep (2016) 35: 1451. doi:10.1007/s00299-016-1987-x

Abstract

CRISPR/Cas9 has evolved and transformed the field of biology at an unprecedented pace. From the initial purpose of introducing a site specific mutation within a genome of choice, this technology has morphed into enabling a wide array of molecular applications, including site-specific transgene insertion and multiplexing for the simultaneous induction of multiple cleavage events. Efficiency, specificity, and flexibility are key attributes that have solidified CRISPR/Cas9 as the genome-editing tool of choice by scientists from all areas of biology. Within the field of plant biology, several CRISPR/Cas9 technologies, developed in other biological systems, have been successfully implemented to probe plant gene function and to modify specific crop traits. It is anticipated that this trend will persist and lead to the development of new applications and modifications of the CRISPR technology, adding to an ever-expanding collection of genome-editing tools. We envision that these tools will bestow plant researchers with new utilities to alter genome complexity, engineer site-specific integration events, control gene expression, generate transgene-free edited crops, and prevent or cure plant viral disease. The successful implementation of such utilities will represent a new frontier in plant biotechnology.

Keywords

Cas9 CRISPR Crop improvement Gene editing Gene knock-in Functional genomics 

Abbreviations

CRISPR

Clustered regularly interspaced short palindromic repeats

Cas9

CRISPR-associated protein 9

dCas9

Dead Cas9

PAM

Protospacer adjacent motif

RNP

Ribonucleoprotein

sgRNA

Single guide RNA

TALEN

Transcription activator-like effector nuclease

Introduction

Among all of the scientific advances from this century, few have displayed the potential to revolutionize biology more than genome-editing. Genome-editing has become possible with the development of molecular technologies, such as RNA/DNA gene repair oligonucleotides (Gamper et al. 2000), zinc finger nucleases (Townsend et al. 2009), TALENs (Bogdanove and Voytas 2011; Christian et al. 2010; Li et al. 2011), homing endonucleases (Hafez and Hausner 2012), and CRISPR/Cas9 (Feng et al. 2013; Jinek et al. 2012; Li et al. 2013; Nekrasov et al. 2013; Shan et al. 2013). Although each method has its advantages and disadvantages, CRISPR/Cas9 has become the system of choice due to its versatility and ease of use (Belhaj et al. 2013, 2015; Bortesi and Fischer 2015; Hsu et al. 2013; Lozano-Juste and Cutler 2014). CRISPR/Cas9 has filled a molecular biology need for a sequence-specific recognition tool as well as a simple technique to produce sequence-specific genomic breaks. With respect to the plant sciences, CRISPR/Cas9 is also important as it provides an avenue to avoid the traditional transgenic classification (Voytas and Gao 2014).

CRISPR/Cas9-based technologies typically utilize two components, CRISPR-associated protein 9 (Cas9) and a single guide RNA (sgRNA), to perform genome-editing and other molecular functions. Originally identified in Streptococcus pyogenes, CRISPR/Cas9-mediated double-strand breakage relies upon two interacting RNA moieties: CRISPR RNAs (crRNA) and trans-activating RNAs (tracrRNA) for sequence specificity (Deltcheva et al. 2011; Garneau et al. 2010; Gasiunas et al. 2012; Wiedenheft et al. 2012). It was shown that a single chimeric RNA composed of these two RNA moieties could be utilized to recruit Cas9 in a site specific manner, increasing the ease in CRISPR/Cas9 use in biotechnological applications (Jinek et al. 2012). The sgRNA typically is designed to contain a 20-nucleotide sequence complementary to a target sequence directly in front of a protospacer adjacent motif (PAM) (Gasiunas et al. 2012; Jinek et al. 2012). The Cas9 can then interact with the sgRNA and target DNA creating a double strand break in the target DNA 3–4 base pairs upstream of the PAM site (Jinek et al. 2012).

Since the discovery of the CRISPR prokaryotic defense system, scientists have exploited properties of this DNA cleavage mechanism to allow directed alterations to be made within the genomes of virtually any organism. As a result, plant scientists have quickly adopted this mutagenesis technology, accelerating insight into plant biological processes and altering plant traits, such as powdery mildew resistance in wheat (Wang et al. 2014) and virus resistance in cucumber (Chandrasekaran et al. 2016). For more information in the use of CRISPR/Cas9 and other genome-editing technologies for crop improvement, the reader is directed to these reviews (Cardi 2016; Schaeffer and Nakata 2015). As the uses of CRISPR/Cas9 technologies improve and evolve, their impact upon the plant sciences continues to increase. Such technologies have included techniques to reduce off-target genomic editing (Fauser et al. 2014), perform multiplexing (Li et al. 2013; Zhou et al. 2014), and directly knock-in gene sequences through homologous recombination (Schiml et al. 2014). The rapid dissemination of CRISPR/Cas9 technologies will endure, especially in light of the continued evolution of this technology across all fields of biology.

In this review, we examine the recent evolution of CRISPR/Cas9-mediated technologies occurring in non-plant systems and speculate on how these technologies may benefit plant functional genomics and plant biotechnology in the near future. New applications and modifications of the CRISPR/Cas9 system are being developed at a swift pace adding to the number of genome-editing tools at the molecular scientist’s disposal. We envision that the next generation of CRISPR/Cas9 tools will lead to an alteration in our perception of what is possible in the future of plant biotechnology. For additional discussions into other aspects of CRISPR/Cas9 and its application in plant biotechnology, the reader is referred to a number of earlier reviews (Belhaj et al. 2013, 2015; Bortesi and Fischer 2015; Chen and Gao 2013; Fichtner et al. 2014; Jain 2015; Kumar and Jain 2015; Mahfouz et al. 2014; Nagamangala Kanchiswamy et al. 2015; Puchta 2016; Raitskin and Patron 2016; Schaeffer and Nakata 2015; Schiml and Puchta 2016; Sprink et al. 2015; Voytas and Gao 2014; Weeks et al. 2015; Zhang et al. 2016a).

Future of CRISPR in plant functional biology and biotechnology

The first applications of CRISPR/Cas9 in plants were reported in August 2013 and involved the generation of site-specific mutagenesis events in Arabidopsis thaliana (Feng et al. 2013; Li et al. 2013), Oryza sativa (Feng et al. 2013; Shan et al. 2013), Nicotiana benthamiana (Li et al. 2013; Nekrasov et al. 2013), and Triticum aestivum (Shan et al. 2013). Since these reports, the number of plant species edited, and techniques to implement CRISPR/Cas9-mediated genome-editing, has grown (Table 1). It is expected that this list will grow further as plant genomic resources, as well as techniques to perform transient or stable studies, are developed for new plant species and cultivars.
Table 1

Current status of CRISPR/Cas9-mediated editing within the Plantae kingdom

Species

Common name

Technique(s) employed

Reference(s)

Arabidopsis thaliana

Thale cress

Agrobacterium-mediated transient expression assays

Jiang et al. (2013)

Agrobacterium-mediated transformation

Fauser et al. (2014)

Protoplast transfection with Cas9/sgRNA ribonucleoproteins

Woo et al. (2015)

Citrus × sinensis

Sweet orange

Xanthomonas citri subsp. citri-facilitated agroinfiltration

Jia and Wang (2014a)

Citrus × paradisi Macf.

Duncan grapefruit

Xanthomonas citri subsp. citri-facilitated agroinfiltration

Jia and Wang (2014b)

Chlamydomonas reinhardtii

Electroporation of cell wall free mutant

Jiang et al. (2014)

Cucumis sativus L.

Cucumber

Agrobacterium-mediated transformation

Chandrasekaran et al. (2016)

Glycine max

Soybean

Hairy root transformation

Jacobs et al. (2015)

Biolistic transformation

Jacobs et al. (2015)

Lactuca

Sativa

Lettuce

Protoplast transfection with Cas9/sgRNA ribonucleoproteins

Woo et al. (2015)

Marchantia polymorpha L.

Liverwort

Agrobacterium-mediated transformation

Sugano et al. (2014)

Nicotiana attenuate

Coyote tobacco

Protoplast transfection with Cas9/sgRNA ribonucleoproteins

Woo et al. (2015)

Nicotiana benthamiana

Agrobacterium–mediated transient expression assays

Nekrasov et al. (2013), Jiang et al. (2013)

HR in transfected protoplasts

Li et al. (2013)

Nicotiana tabacum

Tobacco

Protoplast transfection

Gao et al. (2015)

Agrobacterium-mediated transformation

Gao et al. (2015)

Oryza sativa

Rice

Agrobacterium-mediated transformation

Feng et al. (2013)

Protoplast transfection with plasmid

Jiang et al. (2013), Shan et al. (2013)

Particle bombardment

Shan et al. (2013)

Protoplast transfection with Cas9/sgRNA ribonucleoproteins

Woo et al. (2015)

Petunia × hybrid

Petunia

Agrobacterium-mediated transformation

Zhang et al. (2016b)

Populus tomentosa

Chinese white poplar

Agrobacterium-mediated transformation

Fan et al. (2015)

Populus tremula × alba

Gray poplar

Agrobacterium-mediated transformation

Zhou et al. (2015)

Solanum lycopersicum

Tomato

Agrobacterium-mediated transformation

Brooks et al. (2014)

Solanum tuberosum

Potato

Agrobacterium-mediated transformation

Wang et al. (2015)

Sorghum bicolor

Sorghum

Agrobacterium-mediated transformation

Jiang et al. (2013)

Triticum aestivum

Wheat

Protoplast transfection

Shan et al. (2013), Wang et al. (2014)

Biolistic transformation

Wang et al. (2014)

Zea mays

Corn

Protoplast transfection

Liang et al. (2014)

Agrobacterium-mediated transformation

Feng et al. (2016)

The implementation of CRISPR/Cas9 technologies has rapidly spread through the plant kingdom. Many different transient expression and transformational protocols have been established within these systems

The current state of CRISPR/Cas9 technologies in the biosciences is summarized in Table 2. Many of these technologies have yet to be tested or established in plant systems. Nevertheless, we foresee many of these technologies having useful roles in future studies in plant biology and biotechnology. These technologies include procedures allowing for the insertion or removal of specific DNA sequences within a target genome, the activation or repression of gene expression, and new tools to enable the manipulation of biochemical processes. In addition, the repertoire of Cas9 or Cas9-like proteins is increasing through identifying new variants in alternative species or through rational engineering. Most of these new variants remain untested in plant systems, but possess unique properties, which can overcome some limitations of Cas9.
Table 2

Current state of the art of CRISPR/Cas9 technologies throughout all disciplines

 

Utility

Species first reported

Citation

Employed in plant system?

References

Gene knock in-HR

Precise transgene insertion

Mouse

Yang et al. (2013)

Arabidopsis thaliana

Schiml et al. (2014)

Zea mays

Svitashev et al. (2015)

Gene knock in-NHEJ

Precise transgene insertion

Zebrafish

Kimura et al. (2014)

No

Native gene-GFP fusion

Protein localization

Human cells

Ratz et al. (2015)

No

Split Cas9

Control over editing

Human cells

Wright et al. (2015)

No

Human cells

Zetsche et al. (2015a)

Genome-wide functional screening

sgRNA libraries

Functional gene analysis

Mouse

Koike-Yusa et al. (2014)

No

Human cells

Schmidt et al. (2015)

Repression screening

Functional gene analysis

Human cells

Gilbert et al. (2014)

No

Activation screening

Functional gene analysis

Human cells

Gilbert et al. (2014), Konermann et al. (2015)

No

Enhancer  screening

Identify/characterize regulatory regions

Human cells

Korkmaz et al. (2016)

No

Transgene-free genome editing

Viral-encoded sgRNA/Cas9 editing

Editing multicellular organisms

Human cells

Shalem et al. (2014)

Nicotiana tabacum

Baltes et al. (2014)

Editing with sgRNA/Cas9 RNPs

DNA-free genome editing

Human cells

Kim et al. (2014)

Arabidopsis thaliana

Woo et al. (2015)

Oryza sativa

Woo et al. (2015)

Nicotiana attenuata

Woo et al. (2015)

Lactuca sativa

Woo et al. (2015)

Removal of transgene through crossing/selfing

Transgene-free genome editing

Oryza sativa

Xu et al. (2015)

Changes in chromosome structure

Deletion of large genomic fragments

Functional analysis, trait improvement

Zebrafish

Xiao et al. (2013)

Oryza sativa

Zhou et al. (2014)

Insertion of large fragments

Trait improvement

Pig-9.4 kb

Ruan et al. (2015)

No

Mouse-5.3 kb

Zhang et al. (2015)

Non-homologous chromosome translocation

Directed chromosome rearrangement

Caenorhabditis elegans

Chen et al. (2015)

No

Virus removal/resistance

Cleavage of viral genome

Engineering viral resistance

Human cells

Price et al. (2015), Ramanan et al. (2015)

Nicotiana benthamiana

Baltes et al. (2015), Ji et al. (2015)

Genome “cleaning” of virus

Generation of virus-free material

Human cells

Hu et al. (2014)

No

Pig

Yang et al. (2015)

Cas9/sgRNA RNP removal of virus

Generation of virus-free material

No

RCas9

ssRNA cleavage

Alter transcript abundance, cleave viral RNA

In vitro

O’Connell et al. (2014)

No

mRNA localization

Track mRNA localization and deposition

Human cells

Nelles et al. (2016)

No

dCas9

Nickase

Reduce off-target editing

Human cells

Mali et al. (2013)

Arabidopsis thaliana

Fauser et al. (2014)

Mouse

Ran et al. (2013)

Nuclease fusion

Reduce off-target editing

Human cells

Guilinger et al. (2014), Tsai et al. (2014)

Arabidopsis thaliana

Fauser et al. (2014)

Gene activation

Understand gene function

Human cells

Mali et al. (2013), Gilbert et al. (2013)

N. benthamiana

Piatek et al. (2015)

Yeast

Gilbert et al. (2013)

Gene repression

Understand gene function

Human

Gilbert et al. (2013)

N. benthamiana

Piatek et al. (2015)

Yeast

Gilbert et al. (2013)

Targeting lncRNA (CRISPR-Display)

Study localization of lncRNAs

Human cells

Shechner et al. (2015)

No

Epigenetic regulation

Targeted histone acetylation

Study/alter specific histones

Human cells

Hilton et al. (2015)

No

Targeted histone demethylation

Study/alter specific histones

Human cells

Kearns et al. (2015)

No

The molecular biology applications of CRISPR/Cas9 technologies have rapidly expanded since its first biotech application. Many of these were first demonstrated in non-plant systems, and later implemented within plant systems. Many applications of CRISPR/Cas9 remain untested in plant systems, but will likely become essential tools in plant biotechnology

The future of plant gene discovery, functional genomics, and trait improvement will undoubtedly be influenced by new developments in CRISPR/Cas9 technologies. Recent improvements to Cas9 (Kleinstiver et al. 2016; Slaymaker et al. 2016), discovery of alternatives to SpCas9 (Kleinstiver et al. 2015; Zetsche et al. 2015b) along with the establishment of new CRISPR/Cas9 applications suggest that CRISPR/Cas9 technologies will continue to revolutionize plant biology. These new utilities will likely transform the ability of plant scientists to direct site-specific integration events, edit the epigenome, regulate gene expression, generate transgene-free edited plants, perform large-scale modifications of chromosome structure, edit plants at the multicellular level, and abolish or treat viral disease.

Enabling site-specific DNA integration

One of the more promising roles of CRISPR/Cas9 is its use in gene knock-in or directed gene insertion. This technique was designed to facilitate the insertion, removal, or replacement of a specific gene(s) within a given genome. As a part of this procedure, a single or double genomic break is generated in the host genome. Introduced DNA can then be inserted by the host native DNA repair machinery through either non-homologous end joining (NHEJ) or homologous recombination (HR) pathways, as shown in studies using zebrafish (Auer et al. 2014; Kimura et al. 2014) and Caenorhabditis elegans (Chen et al. 2013). Studies in mammalian cell systems have demonstrated that the efficiency of CRISPR/Cas9-mediated knock-in through HR can be increased through silencing or inhibiting key members of the NHEJ pathway (Chu et al. 2015; Maruyama et al. 2015). Increases in the efficiency of NHEJ-mediated gene knock-in were obtained through the use of a destabilization domain-fused Cas9 variant (Cas9-DD) in human and mouse cells (Geisinger et al. 2016). Efficiency using Cas9-DD was seen to be as high as 36 % and did not utilize selection. Site-specific gene knock-in of selectable markers has already been performed in Arabidopsis (Schiml et al. 2014) and maize (Svitashev et al. 2015), but efficiency of non-selectable marker knock-in in plants has yet to be reported. Gene knock-in is particularly advantageous in plant studies, since it would alleviate technical challenges, such as position effects and copy number variability associated with currently used procedures for DNA integration. In addition, refinement of this technique could allow the assessment of gene functionality. For instance specific genes, SNPs, alleles, regulatory elements, exons, or introns could be either introduced or removed to probe their functions, while retaining the appropriate genomic context.

Gene knock-in or insertion also holds promise in other areas of plant functional biology, such as assessing gene expression or protein localization. Employing these approaches in experimental design may better reflect endogenous expression levels compared with other transgenic methods by maintaining the position, copy number, and epigenetic regulations. Protein fusions with GFP have been generated in human cells through CRISPR/Cas9 mediated gene insertion directly downstream of native genes (Ratz et al. 2015). This technique is powerful as it can enable protein localization studies in the correct genomic context, likely retaining native expression. One could also repurpose this technique through incorporating other functional molecular tools, such as β-glucuronidase (GUS) after native promoters to assess native promoter activity or by changing protein targeting sequences to alter their localization. Plant proteins could also be purified by directly integrating HIS-tag or other tag-encoding sequence in frame with the native gene.

Controlling gene expression

Knowledge of the dynamic and complex processes by which information stored within individual genes and gene networks support biological growth and development is fundamental to our understanding of life. The ability to precisely control the expression of a given gene or set of genes would expedite insights into the function of individual genes and their role in the complex gene networks that regulate development. Previous plant studies relied primarily on RNAi technologies to control gene expression, which utilized small interfering RNAs or short hairpin RNAs (Watson et al. 2005; Small 2007). Although useful, RNAi technologies were frequently found to be inefficient and nonspecific (Jackson et al. 2003). Other targeted methods for gene regulation, such as that utilizing zinc finger or transcription activator-like effectors, can robustly target DNA through programmable DNA binding domains and actively recruit effectors for either the repression or the activation of gene expression (Christian et al. 2010; Gaj et al. 2013; Kabadi and Gersbach 2014; Porteus and Baltimore 2003). DNA binding proteins need to be individually designed, however, which renders their construction and delivery for the regulation of multiple loci a technical limitation. CRISPR technology can be repurposed to contain the convenience and scalability of RNAi with the robustness and modularity of DNA binding proteins (Dominguez et al. 2016).

Transcriptional regulation

The first report of utilizing CRISPR/Cas9 technologies to control gene expression involved the use of a Cas9 variant known as “dead Cas9” (dCas9) to suppress gene expression in E. coli (Qi et al. 2013). The dCas9 is a catalytically inactive form of the nuclease that retains its ability to bind DNA. The use of dCas9 to inhibit gene expression has been referred to as CRISPR interference (CRISPRi). The CRISPRi technology was shown to be efficient, specific, and adaptable for multiplexing (Dominguez et al. 2016). Evidence suggests that CRISPRi exerts its inhibitory function by forming a DNA recognition complex, composed of dCas9 and sgRNA, which specifically interferes with transcription factor binding, transcriptional elongation, or RNA polymerase function.

Targeted activation of gene expression using the CRISPR technology was also shown to be possible in bacteria. Researchers utilized dCas9 fusion proteins to recruit transcriptional activators to the target promoter (Bikard et al. 2013). Since this initial study, a number of reports have shown that this CRISPR activation technology also functions in mammalian cells by measuring the activation of both reporter and endogenous genes (Gilbert et al. 2013; Konermann et al. 2015; Maeder et al. 2013). Gene activation and repression in planta with dCas9-based transcription factors have been reported in Nicotiana benthamiana (Piatek et al. 2015) and A. thaliana (Lowder et al., 2015). Through fine-tuning these tools for other plant systems, plant biologists will be able to utilize a convenient and robust method to alter gene expression and elucidate gene function.

Post-transcriptional regulation

Cas9 can be utilized to specifically bind and/or cleave RNA rather than the usual DNA target (Nelles et al. 2015; O’Connell et al. 2014). This instance of Cas9 use, termed RCas9, relies upon the design and use of a sgRNA and DNA “PAMmer” to target a single strand RNA target sequence (O’Connell et al. 2014). The use of an RNA-specific CRISPR/Cas9 system opens the door to perform many functional studies in RNA biology, including silencing a gene through RNA cleavage, enhancing or blocking translation, altering splice variants, probing RNA localization, or functioning as a molecular tag useful in the isolation of specific RNAs and their interactors (Nelles et al. 2015; O’Connell et al. 2014). Further developments in engineering single-strand and double-strand RNA-specific Cas9 variants are likely to occur in the future.

While still in its infancy, the use of RCas9 holds great promise in plant functional genomic studies. This technology has been proposed as a means to control gene expression (O’Connell et al. 2014) through cleaving specific transcripts and could serve as a viable alternative to antisense, RNA interference (RNAi), or virus-induced gene silencing (VIGS). RCas9 has been used in human cells to bind to mRNA and to display mRNA localization and distribution (Nelles et al. 2016). Through pairing RCas9 with inducible promoters or an inducible Cas9 assembly it may be possible to study embryo lethal plant gene function. RCas9 could also be of interest to those interested in engineering resistance to or removing RNA-genome viruses. The implementation and efficiency of each of these roles of RCas9 still need to be assessed in planta.

dCas9 can be repurposed to direct functional RNA or RNA moieties by incorporating them to the 3′ end of a sgRNA (Shechner et al. 2015). This technology is termed CRISPR-Display or “CRISP-Disp” and has been utilized to localize long non-coding RNAs (lncRNAs) as long as 4.8 kb to specific genomic locations in human cells. It is anticipated that such a technology will empower new research in resolving the function of noncoding RNAs (Lin and Corn 2015; Shechner et al. 2015). CRISP-Disp can also be employed to modulate gene expression if the RNA moiety plays a role in gene expression. In addition to probing the interaction of non-coding RNA with DNA, a similar strategy using RCas9 has been proposed to target and/or investigate the interaction of non-coding RNA with other RNAs (Lin and Corn 2015). The field of plant lncRNAs is very young, but these molecules have already been shown to participate in important plant biological processes, such as photomorphogenesis, fertility, and flowering (Liu et al. 2015). CRISP-Disp will likely facilitate an increase in the understanding of long non-coding RNA roles in plants and possibly serve as a means to alter these traits.

Epigenetic regulation

Epigenomic modifications, such as histone acetylation and methylation or DNA methylation (Jaenisch and Bird 2003; Rice and Allis 2001; Turck and Coupland 2014), can have significant effects on plant gene expression (Chen 2007), transposon silencing (Lisch 2009), and likely other phenomena. Many important plant traits, such as floral symmetry (Cubas et al. 1999), vernilization (Song et al. 2012), and plant stress response (Boyko and Kovalchuk 2008), have been shown to be regulated, at least in part, epigenetically. Efforts to elucidate the mechanisms regulating the epigenome have been hampered by technical limitations in the ability to manipulate chromatin modifications. Such a technical advance would not only expedite investigations to unravel the secrets of the epigenome, but also in the development of strategies to harness such technology for use as a molecular tool.

Recently, it was shown that a dCas9-acytyltransferase fusion protein (dCas9p300 Core) was able to acetylate histone H3 proteins in site-specific loci in human cells (Hilton et al. 2015). This acetylation of histones at promoter and enhancer sites induced target gene expression. In contrast, repression of the Oct4 gene and a loss of pluripotency were observed in embryonic stem cells upon co-expression of a lys-specific histone demethylase 1 (LSD1) and an Oct4 enhancer motif targeting sgRNA (Kearns et al. 2015). In addition, researchers showed that targeting a dCas9-Kruppel-associated box (KRAB) repressor fusion along with a sgRNA targeting the HS2 enhancer element resulted in the silencing of multiple globin genes through H3K9 trimethylation. Overall, these epigenetic studies conducted in mammalian systems demonstrate that the repurposed CRISPR platform provides a useful tool to gain insights into the mechanisms regulating the epigenome. One can envision the use of such a platform in high-throughput screens of regulatory-element function as well as in biotechnological applications. In plants, the use of CRISPR to modulate epigenetic modifiers has also been proposed to control the activation of retrotransposons to enable the increasing genetic diversity for use in breeding crop traits (Paszkowski 2015).

Genome-wide functional screening

Development of genome-scale mutant libraries has facilitated the genetic elucidation of gene function in a variety of organisms. The generation of such libraries has included the use of chemical mutagens, irradiation (Ahloowalia and Maluszynski 2001), or random integration of foreign DNAs (Krysan et al. 1999; Tadege et al. 2008). These techniques, however, can lead to difficult downstream analysis due to the mutation of multiple loci, preference to mutate specific loci, or lengthy processes associated with identifying the mutated locus. CRISPR/Cas9 may be repurposed as a new means to enable high-throughput sequence screens. Large libraries of CRISPR/Cas9 vectors containing sgRNAs have already been developed,which would target nearly every protein-coding gene in human cells (Schmidt et al. 2015) and mouse (Koike-Yusa et al. 2014). Such a resource could be of great use in high throughput forward functional gene screening in plant systems. Non-coding regions, such as enhancer regions, have also been characterized in human cells through designing sgRNA libraries targeting potential enhancer sites (Korkmaz et al. 2016). Performing mutation screens with sgRNA libraries could also make identification of mutated genes simpler than other methods. Although random mutant plant populations have been generated utilizing EMS, T-DNA, and transposons, the development of additional genome-scale mutant populations would still be of benefit. This approach would also gain from Cas9’s ability to cleave methylated DNA (Hsu et al. 2013). In addition, as CRISPR/Cas9 can generate bi-allelic mutations in T0 plants (Gao et al. 2015; Ma et al. 2015; Zhang et al. 2014), it is possible that seed derived from edited plants could be directly utilized to observe phenotypes. As the transformation cassette and sequence of the sgRNA will be retained in stably transformed mutant plants, it could be identified and used to scan the genome to find putative cleavage sites. These libraries can also be used in conjunction with alternative forms of Cas9, such as dead Cas9 (dCas9) to enhance or disrupt gene expression. As targeted activation of retrotransposition could potentially be used for generating phenotypic diversity for applications in plant breeding (Paszkowski 2015), a Cas9-based transposon activation strategy could also be utilized to generate mutant libraries for high-throughput gene screening.

In addition to a mutation-based approach, CRISPR/Cas9 technologies can be utilized to screen for phenotypes through genome wide-activation or repression screens. Activation screens have already been demonstrated with CRISPR/Cas9 in human cell systems (Gilbert et al. 2014; Konermann et al. 2015) and successfully identified genes that, when upregulated, conferred altered phenotypes, such as resistance to toxins or inhibitors. Similarly, repression screens have been successfully employed in human cells (Gilbert et al. 2014). Employment of such techniques can provide a broader understanding of gene function since the resulting phenotypes would be derived from modulations in gene expression rather than a knockout of expression.

Transgene-free modification of plant genomes

Several different transgene-free methods have been utilized, or are now feasible, to perform CRISPR/Cas9-mediated genome-editing in plants (Fig. 1). These approaches often are used for proof-of-concept studies of CRISPR/Cas9 components and vectors in new plant systems or to assess changes in editing efficiencies due to different sgRNA sequences, regulatory elements, or Cas9 variants. Methods to perform these transient studies in plant systems have included Agrobacterium-mediated infiltration (Li et al. 2013), particle bombardment (Miao et al. 2013), and protoplast transfection (Shan et al. 2013) typically using plasmids encoding Cas9 and sgRNAs (Fig. 1B). Such transient methods have been proposed as a non-transgenic means to generate genome-edited plants assuming that no DNA is integrated into the plant genome (Chen and Gao 2013; Podevin et al. 2013). Infection with engineered plant virus encoding CRISPR/Cas9 components (Fig. 1A) also holds potential for a transgene free strategy assuming that the virus can be removed and that engineered viruses are small enough in size to retain the ability to move between cells. Transgene-free genome-editing can alternatively be performed through plant transformation followed by crossing or selfing to “breed out” a transgene (Podevin et al. 2013; Xu et al. 2015).
Fig. 1

Non-transgenic methods for performing plant genome editing. Several methods have been developed which enable or have demonstrated the possibility of editing plant genomes without stable plant transformation. These include (A) the introduction of an engineered virus containing the CRISPR/Cas9 components able to move cell-to-cell through the plasmodesmata, (B) transient introduction of vectors encoding CRISPR/Cas9 components, and (C) application of Cas9/sgRNA ribonucleoprotein complexes. CRISPR/Cas components depicted include sgRNA (red) and Cas9 (light blue) (color figure online)

A completely DNA-free method for inducing genome modifications was recently demonstrated in Arabidopsis, rice, tobacco, and lettuce (Woo et al. 2015). This technique utilizes Cas9 and sgRNA ribonucleoprotein (RNP) complexes to transfect plant cells (Fig. 1C). Whole plants are then regenerated from transfected and edited cells. The use of such a technique could be important for generating genome-editing in industrial applications as this technique eliminates the possibility of the integration of DNA vectors or transgenic material. The use of RNPs can also simplify the design of experiments, as promoters driving the expression of sgRNAs and codon-usage of the Cas9 gene would not need to be addressed. In addition, the authors of the study suggested that such a method could reduce off-target editing due to a limited exposure to Cas9/sgRNA compared with other methods. One limitation to this technique is the absence of selectable markers; however, analysis of a large number of resultant plants could be sufficient to identify plants with desired mutations.

Generation of cisgenic modifications

The manner in which a plant is modified has become an area of interest due to public perceptions toward the use of genetically engineered plants (Hou et al. 2014). Thus, efforts to develop cisgenic technologies (engineered modifications accomplished through the introduction of DNA sequences only derived from the same or compatible plant species) have been ongoing. Genome-editing alone, or in conjunction with cisgenics, is a promising area for the future of crop breeding (Cardi 2016). Traditionally, Agrobacterium-mediated transformation is used to generate stable transgenic plants. Even if constructs used in Agrobacterium-mediated transformation contain only plant genes and regulatory sequences, resultant plants can contain bacterial T-DNA borders and transgenes insertion occurs randomly in the genome (Schouten et al. 2006) potentially disrupting essential gene function at the site of integration. The CRISPR/Cas9 site-specific gene knock-in or replacement (Auer et al. 2014; Kimura et al. 2014) mechanism has advantages over the Agrobacterium-mediated mechanism of DNA insertion. The CRISPR system utilizes site-directed cleavage and the native cellular homologous recombination system to recombine foreign DNA flanked by homologous sequences, which enables the control over transgene copy number, the potential to introduce only cisgenic material, the performance of site-specific integration events, and the removal of native genes or regulatory elements and replacement with alternative sequences from closely related species. This process has already been used in planta to introduce selectable marker genes at targeted loci (Schiml et al. 2014; Svitashev et al. 2015), however, the transition to this technology from Agrobacterium-mediated transformation has not occurred in most research programs. Improvements in the efficiency of CRISPR/Cas9-mediated gene knock-in within other systems, such as through the use of Cas9-DD (Geisinger et al. 2016), hold promise that gene knock-in of DNA in plants could become routinely performed without selectable markers.

Editing multicellular organisms in vivo

Outside of the plant sciences, particularly in the medical sciences, CRISPR/Cas9 has been touted as a means to edit genetic disorders. CRISPR/Cas9 has been used to correct genetic diseases, such as a dominant allele that causes cataracts (Wu et al. 2013) and hereditary tyrosinemia (Yin et al. 2014) in mice, but these studies are typically utilizing single cell systems, altering organisms at the zygote stage, or editing only a small subpopulation of cells. Larger populations of cells can be edited in vivo through viral transfection (Ding et al. 2014; Swiech et al. 2015), but performing editing in every cell within a system remains a challenge.

Due to the ability of some plant viruses to pass from cell to cell through plasmodesmata (Benitez-Alfonso et al. 2010), it is feasible that the use of a plant viral delivery system could enable the direct editing of a high proportion of cells in a plant (Fig. 1A). Previous efforts have shown that the genome of geminivirus can be engineered with Cas9 and sgRNAs to perform genome-editing in plants (Baltes et al. 2014). However, due to the large size of the Cas9 protein, such engineered viral genomes would likely fall over the size limit required to pass through the plasmodesmata. Therefore, plants must be generated through regeneration of tissues exposed to the engineered virus. To overcome the size limitation of genetic payloads delivered by virus, Cas9-overexpressing plants have been successfully edited with tobacco rattle virus engineered to contain sgRNAs (Ali et al. 2015a). The identification of smaller Cas9 alternatives (Ran et al. 2015) or the use of split Cas systems could also alleviate the size limitation of a virus-based approach.

Interestingly, Cas9-overexpressing plants exposed to a virus engineered to encode sgRNA possessed edited genomes in somatic cells and in their progeny (Ali et al. 2015a). This suggests that such a strategy could be utilized as a simple and quick technique to generate plants with desired mutations or with desired gene insertion events. These editing events would remove the need to use tissue culture or Agrobacterium-mediated transformation and could enable editing in systems traditionally known as recalcitrant to genome engineering, such as Phaseolus vulgaris (Hnatuszko-Konka et al. 2014). Perhaps such modifications do not even need a virus if sgRNA can be transmitted to gametic cells through an alternative means, such as infiltration or direct injection of Cas9-everexpressing plants.

Generation of virus-free plants

As CRISPR/Cas9 enables bacteria to defend against invading phage, it serves as a great candidate to engineer virus resistance in plant systems. Strategies for engineering resistance to plant viruses and removing plant viruses from plant material using CRISPR/Cas9 are shown in Fig. 2. A proof of concept for engineering increased geminivirus resistance has already been demonstrated using Nicotiana benthamiana (Ali et al. 2015b; Baltes et al. 2015; Ji et al. 2015) and Arabidopsis (Ji et al. 2015). Outside of the plant sciences, CRISPR/Cas9 has been used to remove HIV-1 integrated into human cells (Hu et al. 2014). It has also been used to “virally cleanse” viruses integrated into pig genomic DNA in hopes of preventing transmission in cross-species organ transplants (Yang et al. 2015). Implementation of such techniques to “virally cleanse” plant genomes may be of use in the plant sciences for those viruses able to integrate into plant genomes (Harper et al. 2002) or to inactivate plant retrotransposons. Utilizing DNA-free Cas9/sgRNA ribonucleoprotein treatment (Woo et al. 2015) for virus, viroid, and virion removal may serve as an alternative to the current non-transgenic means of meristem culture and heat treatment for virus removal (Gergerich et al. 2015; Houten et al. 1968).
Fig. 2

Using CRISPR to treat viral disease and engineer resistance to plant viruses. Several studies in plant and other systems have demonstrated the roles that CRISPR may be used in virus treatment and resistance. Virus-free plant material could be obtained through treatment with Cas9/sgRNA RNPs or vectors encoding CRISPR/Cas9 components. Alternatively, plant genomes can be engineered to encode CRISPR/Cas9 components to incorporate a permanent mechanism of viral resistance. As plant viruses can have different genome types, Cas9 could be used to target either ssDNA or dsDNA (1). Alternatively, an RCas9 approach or development of a new variant of Cas9 able to target RNA could be used to target viral RNA genomes (2). CRISPR/Cas9 can also be utilized to remove virus stably integrated into the genome (3). CRISPR/Cas components depicted include sgRNA (red) and Cas9 (light blue) (color figure online)

The recent demonstrations of engineering plant virus resistance are reliant upon human design of sgRNA targeting a viral genome (Ali et al. 2015b; Baltes et al. 2015; Ji et al. 2015). Such a strategy is likely not sufficient to engineer viral resistance for plants living outside of the laboratory due to the number of different pathogens and ever changing genomic sequences of viruses. While still not completely understood, bacteria are able to integrate viral DNA into their own genomes to serve as a means to prime cleavage of foreign DNA with CRISPR-Cas9 (Bolotin et al. 2005; Mojica et al. 2005). New advances have begun to characterize bacterial components involved in this recognition and integration of foreign DNA (Nuñez et al. 2015). As this process becomes better characterized, it opens up the possibility of engineering plants with not only resistance to viruses, but also a means to adapt to a wide range of different viruses and an evolving barrage of viruses encountered in the environment. Furthermore, complementing such a strategy with a Cas9 variant able to cleave RNA could potentially broaden the effectiveness of engineering viral resistance against a wider range of viruses, such as those with ssRNA or dsRNA genomes.

While the potential of utilizing CRISPR/Cas9 as a new strategy to prevent or treat viral disease in plants has been shown, several limitations have been hypothesized that may hinder its implementation. Such proposed limitations include the accumulation of off-target editing events within the plant genome, especially in instances where the editing components are constitutively expressed (Zaidi et al. 2016). To minimize off-target editing of the plant genome, it is possible that alternative and more precise technologies, such as double nickase (Mali et al. 2013), nuclease-fusion (Guilinger et al. 2014), or a Cas9 variant with high fidelity could be employed. Another approach would be to control the expression of CRISPR/Cas9 components through virus-inducible promoters, as has been demonstrated in an insect system (Dong et al. 2016). A second foreseeable problem is a continued evolution of viral genomic sequences targeted by a sgRNA (Zaidi et al. 2016). It is feasible that the sequence-specificity of sgRNA could favor the development of mutations at the sgRNA complementary site in a viral genome. In studies looking to use CRISPR/Cas9 to cleave HIV-1genomic sequence, it was observed that while viral genomes were cleaved, some were repaired with NHEJ with mutations and became immune to further CRISPR/Cas9 cleavage (Wang et al. 2016). If this occurred in plant systems, viruses could then evade the engineered CRISPR/Cas9-mediated resistance in the plant. It is possible that such evolution can be minimized by targeting conserved sequences, engineering multiple sgRNAs targeting a genome, or through introducing new germplasm with sgRNAs targeting the newly evolved viral genomes.

Manipulating chromosome number/structure

The development of double haploid plant germplasm has become an increasingly important component of modern breeding programs. Such resources enable simpler genomes for functional genomic studies and also serve as pure inbred lines that can be used as material for generating hybrids. Currently, the development of such material is limited to techniques, such as gametic cell rescue through tissue culture (Guha and Maheshwari 1964) or interspecific crossing (Bains and Howard 1950; Dunwell 2010). CRISPR/Cas9 has great potential to enable a controlled alteration in chromosome number in plants. In plants, haploidization can be induced through a single point mutation in the CENH3 gene (Karimi-Ashtiyani et al. 2015). Recreation of this point mutation could be performed across economically important crop systems using CRISPR/Cas9. Such an alteration could lead to an easier method to develop haploid germplasm.

In mouse, it has been shown that, despite the tightly packed nature of centromeric DNA, a GFP–dCas9 fusion protein can be directed to bind and label centromeric regions (Anton et al. 2014). As this would likely hold true for plant systems, CRISPR/Cas9 could be potentially repurposed to bind plant centromeric regions. Such binding could be used to disrupt kinetochore assembly and potentially be used to selectively alter chromosome number to induce polyploidy or double haploidy. Such a method could be further modified by employing the DNA-free Cas9/sgRNA RNPs, so that resultant plants lack mutations, maintaining functional mitosis and meiosis.

CRISPR/Cas9 will likely play an important role in performing large-scale alterations in plant genomes. The removal of large genomic regions through the use of sgRNA multiplexing has been employed to delete large genomic fragments (>100 kb) in rice (Zhou et al. 2014). Such a technique could prove valuable in removing multiple plant genes involved in a biosynthetic pathway or elucidating gene function. Insertion of large genomic fragments in plants could also be achieved through CRISPR/Cas9-mediated gene insertion. Currently, the largest reported directed gene insertion using CRISPR/Cas9 in plants is a 2.1 kb nptII gene (Schiml et al. 2014). Larger insertions have been reported in mouse (5.3 kb) (Zhang et al. 2015) and in pig (9.4 kb) (Ruan et al. 2015). The limit on the size of directed gene insertion using CRISPR/Cas9 is yet to be resolved, but consecutive insertion events could be employed to address size limitations.

In addition to removing or adding large genomic fragments, CRISPR/Cas9 holds promise for altering chromosome structure. It has been shown in C. elegans that the CRISPR/Cas9-induced cleavage of two non-homologous sites can induce reciprocal chromosome fragment translocation (Chen et al. 2015). This suggests that the directed exchange of large chromosomal fragments between non-homologous chromosomes is also possible in plant systems. Such a technique could be a powerful tool for plant breeding programs and potentially utilized to introduce synthetic chromosome fragments or pieces of chromosomes from non-sexually compatible organisms in a directed manner.

Alternatives and improvements to SpCas9

While Cas9 has transformed the plant sciences, it possesses many undesirable features that limit its application. These issues include the large size of Cas9; the requirement of a PAM site for cleavage; the generation of blunt ends after cleavage; and promiscuous off-target cleavage. Most studies using CRISPR/Cas9 have utilized the Cas9 from Streptococcus pyogenes (SpCas9) for genome-editing. To address undesirable attributes or alter its traits, modifications to SpCas9 have been made, and the search for SpCas9 alternatives has intensified in recent years (Table 3).
Table 3

Alternatives to S. pyogenes Cas9

Gene/variant

Species of origin

Special attributes

Species tested

References

Cas9n (nickase)

Engineered variant of S. pyogenes

Generation of single strand break

Arabidopsis thaliana

Fauser et al. (2014)

NmCas9

N. meningitides

Alternative PAM site

Longer PAM sequence

Human cells

Hou et al. (2013)

Reduced off-target activity

 

Lee et al. (2016)

St1Cas9

S. thermophilus

Shorter gene sequence

Bacteria

Human cells

Kleinstiver et al. (2015)

SaCas9

S. aureus

Shorter gene sequence

Bacteria

Human cells

Kleinstiver et al. (2015)

eSpCas9

Engineered variant of S. pyogenes

Reduced off-target editing

Human cells

Slaymaker et al. (2016)

SpCas9-HF1

Engineered variant of S. pyogenes

Reduced off-target editing

Human cells

Kleinstiver et al. (2016)

Cas9-DD

Destabilized Cas9

Increase NHEJ-mediated gene insertion efficiency

Human cells

Mouse cells

Geisinger et al. (2016)

Cpf1

F. novicida U112

Alternative PAM site

Staggered target cleavage

Shorter gene sequence

Human cells

Zetsche et al. (2015b)

dCas9p300 Core

acetyltransferase fusion

S. pyogenes

H. sapiens

Targeted histone acetylation

Human cells

Hilton et al. (2015)

To overcome the limitations or requirements for implementation of the CRISPR/Cas9 system, several modifications to the S. pyogenes Cas9 have been made. In adddition, many other organisms possess alternative class 2 CRISPR/Cas9 endonucleases with unique properties. Many of these endonucleases have attributes that will be useful in plant genome editing

Several modifications have been made to the SpCas9 that have improved its application to genome engineering. Rational engineering of Cas9 has been performed to generate enhanced specificity Cas9 (eSpCas9) (Slaymaker et al. 2016) and high fidelity Cas9 (SpCas9-HF1) (Kleinstiver et al. 2016), both of which have shown to have reduced off-target editing. Codon optimization of Cas9 has been performed in efforts to enhance translation, potentially increasing the efficiency of genome-editing. In plant systems, these include codon-optimized variants specifically for Arabidopsis thaliana (AteCas9) (Fauser et al. 2014), Oryza sativa (Shan et al. 2013), Chlamydomonas reinhardtii (Jiang et al. 2014), as well as variants designed for broader groups of organisms, such as the monocot Cas9 (Jiang et al. 2013) and two codon-optimized Cas9 for plants in general (pcoCas9) (Li et al. 2013) (Cas9p) (Ma et al. 2015). Plant codon optimization of high fidelity variants of SpCas9 will likely benefit plant biology researchers in the near future.

Cas9 homologs from bacteria other than S. pyogenes have also been targeted in an effort to find new or alternative attributes. Neisseria meningitidis possess a Cas9 homolog (NmCas9) with an alternative PAM site requirement of 5′-NNNNGATT-3′ and has been successfully used for genome-editing in human cells (Hou et al. 2013). NmCas9 has reduced editing efficiency compared with SpCas9, but displays lower off-target editing (Lee et al. 2016). The PAM site of Streptococcus thermophilus Cas9 (5′-NNAGAAW-3′) also differs greatly from SpCas9 (Garneau et al. 2010) and could enable further potential genomic targets for editing. Engineering of Cas9 has also enabled the engineering of alternative PAM sites (Kleinstiver et al. 2015).

One approach to overcome the negative aspects of Cas9 is through using split-Cas9 architecture. Work has already been done to demonstrate that splitting the Cas9 into two separate components can yield a functional Cas9 able to cleave target sequences (Wright et al. 2015; Zetsche et al. 2015a). Separation of Cas9 into smaller fragments could allow the virus-based technique to employ CRISPR/Cas9 as movement of viral genomes from cell to cell is limited by size and, therefore, cargo size. Controlling the expression of the Cas9 subunits by different promoters could be used to limited exposure of host genomic DNA and reduce off-target effects. Split-Cas9 has also been engineered to assemble in the presence of rapamycin (Zetsche et al. 2015a), which could pose as another useful tool to control the duration of Cas9 activity in efforts to reduce off-target editing.

Recently, a type V CRISPR/Cas9 CPF gene was identified in Francisella novicida U112 (Zetsche et al. 2015b). Like Cas9, this gene encodes for a class 2 CRISPR/Cas endonuclease, but possesses significantly different attributes (Fig. 3). While native Cas9 utilizes two RNA molecules to guide its specificity (crRNA and tracrRNA), Cpf1 utilizes a single RNA (crRNA) for locus targeting. The PAM site requirement of Cpf1 is T-rich, while that of Cas9 is comprised of NGG or NAG sequences (Gasiunas et al. 2012; Jinek et al. 2012). Cas9 cleaves DNA 3-bp upstream of the PAM site and generates two blunt ends (Gasiunas et al. 2012; Jinek et al. 2012), while Cpf1 cuts downstream of the PAM site in a staggered manner resulting in “sticky ends” (Zetsche et al. 2015b). The generation of these sticky ends could be a trait that establishes Cpf1 as an even more robust tool then Cas9. Zhetsche et al. suggested that exploiting the use of sticky ends could be utilized as a technique to direct DNA into site-specific loci through NHEJ and could offer a more efficient gene knock-in technique in cells with low rates of homology directed repair (HDR), such as non-dividing cells. In plant systems, directed gene insertion has been reported through HDR-mediated insertion of selectable markers, such as nptII in Arabidopsis (Schiml et al. 2014) and MoPAT in maize (Svitashev et al. 2015). With the development of more efficient directed gene insertion systems in plants, it is possible that the directed insertion of genes could become a powerful tool and not confined to genes or cassettes with selectable markers.
Fig. 3

Comparison of Cas9 and Cpf1 characteristics. Cpf1 is a class 2 CRISPR/Cas endonuclease, which possesses alternative traits and cleavage requirements to SpCas9. These traits may enable new and enhance existing CRISPR/Cas technologies in the plant sciences (color figure online)

With the discovery and engineering of alternatives to SpCas9, several limitations to SpCas9 can be overcome. These SpCas9 alternatives could open up new plant genomic target sites for editing, may offer reduced off-target editing or increased editing efficiency, and may be more effective than SpCas9 in other CRISPR/Cas9 technologies. Potentially new applications for genome-editing in plants are going to become available with SpCas9 alternatives.

Concluding remarks

Although CRISPR-mediated genome-editing can be considered a relatively new technology, it has already garnered widespread adoption by life scientists. It appears that CRISPR/Cas9 technologies will continue to advance our understanding of metabolic processes, and holds the promise to transform all areas of biology. In plants, CRISPR/Cas9-mediated genome-editing has been implemented in studies of plant functional genomics and plant trait improvement. As modifications to the CRISPR/Cas9 system continue to evolve and provide new applications, plant scientists will continue to adopt these technologies and refine their use in plant systems. It is anticipated that CRISPR technologies will be a major driving force in the genetic engineering of plants for trait improvement and synthetic biology applications. These emerging technologies enable targeted modification of diverse crop tissues and systems. CRISPR-mediated enhancements in crop yield, shelf life, nutritional content, physical appeal, production of specialty chemicals, and abiotic and biotic stress tolerance all appear to be on the horizon.

Author contribution statement

Both authors wrote and approved the final manuscript.

Acknowledgments

The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. The authors apologize to those whose work could not be cited due to space limitation. The authors would like to thank Dr. Michele McConn, Dr. Christopher Hendrickson, and Ms. Vandhana Krishnan for critically reading this manuscript and Mr. Adam Gillum for his assistance in preparing Fig. 3.

Compliance with ethical standards

Funding

This work was supported by the US Department of Agriculture, Agricultural Research Service, under Cooperative Agreement Number 58-3092-5-001.

Conflict of interest

The authors declare that they have no conflict of interest.

Copyright information

© Springer-Verlag Berlin Heidelberg (outside the USA) 2016

Authors and Affiliations

  1. 1.Department of Pediatrics, Baylor College of MedicineUSDA/ARS Children’s Nutrition Research CenterHoustonUSA