Arabidopsis ALF4 encodes a nuclear-localized protein required for lateral root formation


For correspondence (fax +1 617 353 6340; e-mail

Current address: Biology Department, University of Massachusetts, Amherst, MA 01003, USA.


Lateral root formation, the primary way plants increase their root mass, displays developmental plasticity in response to environmental changes. The aberrant lateral root formation (alf)4-1 mutation blocks the initiation of lateral roots, thus greatly altering root system architecture. We have positionally cloned the ALF4 gene and have further characterized its phenotype. The encoded ALF4 protein is conserved among plants and has no similarities to proteins from other kingdoms. The gene is present in a single copy in Arabidopsis. Using translational reporters for ALF4 gene expression, we have determined that the ALF4 protein is nuclear localized and that the gene is expressed in most plant tissues; however, ALF4 expression and ALF4's subcellular location are not regulated by auxin. These findings taken together with further genetic and phenotypic characterization of the alf4-1 mutant suggest that ALF4 functions independent from auxin signaling and instead functions in maintaining the pericycle in the mitotically competent state needed for lateral root formation. Our results provide genetic evidence that the pericycle shares properties with meristems and that this tissue plays a central role in creating the developmental plasticity needed for root system development.


Plant tissues retain remarkable developmental plasticity, which enables the plant to alter its structure in response to diverse environmental conditions. This plasticity is especially obvious in the architecture of the root where the extent of root branching, or lateral root formation, is dramatically altered by environmental conditions. Lateral root formation mirrors primary root development; a lateral root develops all of the same tissues as the primary root (Dolan et al., 1993). The only apparent difference is that lateral root formation involves the re-initiation of the root tissue program from differentiated pericycle cells present in the primary root or previously formed lateral roots. While lateral root formation is clearly guided by an intrinsic developmental program, environmental factors such as nutrient availability can modulate the program (Fitter et al., 1988). For example, lateral root formation is highly responsive to nitrate concentration (Zhang and Forde, 2000; Zhang et al., 1999) as well as the ratio of sucrose to nitrogen (Malamy and Ryan, 2001).

The number of lateral roots can be altered by either the application of auxin or perturbation of internal auxin levels (Blakely and Evans, 1979; Blakely et al., 1982, 1988; Kares et al., 1990; Klee et al., 1987; Torrey, 1950). Auxin transported from the shoot is necessary for lateral root formation in young seedlings (Casimiro et al., 2001; Reed et al., 1998). In addition, the number of lateral roots is altered in mutants that affect auxin metabolism (Boerjan et al., 1995; Celenza et al., 1995; Delarue et al., 1998; King et al., 1995), transport (Marchant et al., 2002), or signaling (Hobbie and Estelle, 1995; Leyser et al., 1996). Furthermore, a high sucrose/nitrogen ratio, which represses lateral root initiation, alters auxin-responsive gene expression, suggesting that nutrient regulation of lateral root formation is mediated by auxin localization (Malamy and Ryan, 2001).

Lateral root development has been divided into four steps: (i) stimulation and de-differentiation of pericycle cells; (ii) ordered cell divisions and re-differentiation to generate a highly organized lateral root primordium that may include a group of cells functioning as an apical meristem; (iii) emergence of the lateral root primordium via cell expansion; and (iv) activation of the lateral root meristem to permit continued growth of the organized lateral root (Malamy and Benfey, 1997a,b). Only pericycle files adjacent to the two xylem poles of the Arabidopsis root are able to form lateral roots. At each xylem pole there can be one to three pericycle files that will contribute to a lateral root, thereby creating the bipolar pattern of lateral roots seen in Arabidopsis (Dubrovsky et al., 2001).

Increased cell division in the xylem-adjacent pericycle, the first visible step in lateral root formation, is clearly a crucial step of the process. However, activation of the cell cycle is not sufficient to induce lateral roots to form. Overexpression of cyclins leads to an increased growth rate without an increase in the initiation of lateral roots (Cockcroft et al., 2000; Doerner et al., 1996). These findings suggest that an underlying developmental framework is in place and that cell division, while limiting for the rate of growth, is not the trigger for lateral root development. Nevertheless, competency for cell division is a requirement for lateral roots to initiate (Beeckman et al., 2001; Dubrovsky et al., 2000).

Most mutations that alter the root development program would be expected to affect the development of both the primary and lateral roots. One class of mutants specific to lateral root development would be those that affect production or perception of regulatory signals for lateral root formation. For example, the rooty mutant causes increased indole-3-acetic acid (IAA) and results in a dramatic increase in lateral root initiation (Boerjan et al., 1995; Celenza et al., 1995; King et al., 1995). In addition, a mutation in the Aux/IAA gene solitary-root (SLR)-1 does not affect primary root development but blocks lateral root formation at a very early step (Fukaki et al., 2002). The behavior of this mutant supports the intrinsic role of auxin signaling in lateral root formation.

Although specific examples have not been described previously, another class of mutants specific to lateral roots would be those that are unable to carry out the cell division and/or differentiation programs required for primordium formation upon receipt of inductive signals. This idea imbues the xylem-adjacent pericycle with the developmental plasticity found typically in meristem cells and is supported by findings indicating that the xylem-adjacent pericycle, while considered a differentiated tissue, displays certain properties that distinguish it from the other pericycle cell files (Beeckman et al., 2001; Dubrovsky et al., 2000). These properties, such as competence for cell division and the ability to become founder cells for lateral roots, must be maintained by the xylem-adjacent pericycle for long periods subsequent to their formation. Mutants that alter the ability to maintain this potential for cell division and limited re-differentiation would lose the ability to alter root system architecture in response to intrinsic and extrinsic signals.

ALF4 is a prime candidate for a gene involved in lateral root initiation and not in primary root formation. The alf4-1 mutant forms a primary root, but is blocked at a very early stage in lateral root formation. The alf4-1 primary root displays normal gravitropism, suggesting that alf4-1 is not blocked in all auxin-mediated events (Celenza et al., 1995). As exogenous auxin does not appreciably induce lateral roots in the alf4-1 mutant, ALF4 appears to be required at a step downstream of auxin synthesis and transport (Celenza et al., 1995). The alf4-1 mutant is also pleiotropic: it affects shoot development and is male sterile (Celenza et al., 1995).

Here, we describe the cloning of the ALF4 gene and the use of the cloned gene to define the molecular basis for the inability of alf4 mutants to form lateral roots. Using gene expression reporters, we find that ALF4 is expressed in most tissues and that the ALF4 protein appears to be nuclear localized. We also find that auxin does not affect ALF4 levels, intracellular location, or tissue distribution. Our experiments indicate that ALF4 acts downstream and possibly independent of the auxin signal. alf4-1 pericycle cells are impeded in completing mitosis, thus preventing lateral root development. These experiments suggest that ALF4 maintains the xylem-adjacent pericycle in a mitotically active state. Taken together, these results suggest a model in which ALF4 is required to maintain the developmental plasticity of the pericycle so that this tissue can give rise to lateral roots in response to intrinsic or extrinsic signals.


Cloning of the ALF4 gene

Using a map-based approach we cloned the ALF4 gene, which we had previously located on Arabidopsis chromosome 5 in a 2-cM interval bordered by nga249 and m224 (Celenza et al., 1995). We walked to alf4-1 from the these markers using genomic clones contained in yeast artificial chromosome (YAC) or lambda vectors. We used DNA sequence comparison between mutant and the wild type to determine that alf4-1 is a 12-bp deletion in At5g11030, for which no other mutations have been described previously. This assignment was verified by complementation of the alf4-1 mutant with a genomic clone containing only At5g11030 and further confirmed by demonstrating that a similar clone containing an in vitro-made alf4-1 allele failed to complement the alf4-1 mutant. Details of the walk and complementation are contained in Supplementary Material.

By comparison of the ALF4 genomic sequence to the sequences of cDNAs and cloned RT-PCR products, as well as publicly available expressed sequence tags (ESTs; T21771, T76605, AI993277, AV531796, AA395683, and AA404902), we derived a consensus ALF4 cDNA. Our predicted gene structure for At5g11030 is in agreement with the structure released 5/13/03 (

Based on our predicted ALF4 gene structure, the 12-bp deletion found in the alf4-1 mutant removes the junction between the first intron and the second exon of ALF4 (Figure 1a). RT-PCR products amplified from alf4-1 mRNA identified two classes of mis-spliced ALF4 transcripts each of which would result in a severely truncated protein if translated (Figure 1b). Further evidence supporting the assignment of At5g11030 as ALF4 has come from the identification of a T-DNA insertion allele of ALF4 (alf4-063) from the signal collection ( alf4-063 plants have phenotypes identical to those of alf4-1 and contain a T-DNA insertion in the fifth exon of ALF4 1400 bases downstream of the predicted ALF4 initiation codon (Figure 1a).

Figure 1.

Structure of the ALF4 gene, location of alf4 mutations, and identification of aberrantly spliced alf4-1 mutant mRNAs. Exons (boxes) and introns (line with V underneath) are indicated.

(a) Genomic structure of ALF4 and location of the alf4-1 and alf4-063 mutations. The deleted region is indicated for alf4-1, and intron sequences are in lower case; exon sequences are in upper case. Numbering is relative to the ALF4 start codon (+1).

(b) Structure of aberrantly spliced alf4-1 mRNAs and sequence of the predicted alf4-1 mutant protein. The schematic shows the structure of the 5′ end of wild-type ALF4 gene and the mis-spliced first introns generated from alf4-1. The location of the stop codon (UAA) predicted in the alf4-1 mutant mRNAs is indicated. Also shown is the alf4-1 mutant protein. aa encoded only in the alf4-1 mutant are in bold.

ALF4 encodes a protein unique to Arabidopsis with no similarities to known proteins

ALF4 is predicted to encode a 626 aa protein (molecular weight of 69.2 kDa; Figure 2). Although no ALF4-related genes exist in Arabidopsis, tblastn analysis revealed the presence of ESTs encoding ALF4-related proteins in soybean, wheat, rice, barley, and the moss Physcomitrella patens (Table 1). While the presence of these ESTs from other species indicates that ALF4 is conserved in other plant species, no conserved protein domains were found between ALF4 and proteins from other kingdoms.

Figure 2.

Alternatively spliced forms of ALF4 encode 626- or 246-aa proteins.

The 626-aa form is indicated as ALF4, while the 246-aa form is labeled as ALF4ΔC. (*) indicates the translational stop.

Table 1.  GenBank ESTs from other plant species encoding proteins similar to ALF4
Accession numberRegion of similarity in ALF4aSpeciesTissue source
  • a

    Region in ALF4 protein sequence of aa similarity with an EST; result of tblastn analysis using the 626-aa ALF4 protein sequence as a query against the GenBank EST database, using the default BLOSUM62 algorithm. All sequences listed had an E-value < 3e−11.

BI699008 388–487; 508–541SoybeanRoots
BI698991 388–546SoybeanRoots
AI496608 504–624SoybeanRoots
AW351123 504–624Soybean14-day leaf
BI777722 330–447BarleyDrought-stressed root
BI805850 527–624RiceStem
BE470560 66–202Wheat14-day leaf
BG906027 98–269WheatRoots of etiolated seedlings
BE585510 114–336WheatFusarium infected spikes
BE403433 289–478WheatDrought-stressed seedling
BG906028 502–624WheatRust-infected leaf
BJ170878 430–618 Physcomitrella patens Chloronemata, caulonemata and malformed buds
BJ177953 64–191 Physcomitrella patens Chloronemata, caulonemata and malformed buds

From three independent cDNA sources, (a λ YES library (Elledge et al., 1991), EST T21771, and cloned RT-PCR products), we found an alternate splice of intron 5 that results in four bases added to the 3′ end of exon 5. These cDNAs, which represent almost one-quarter of the ALF4 cDNAs identified, result in a stop codon after aa 246 (Figure 2). The 246-aa form of ALF4 (called ALF4ΔC to distinguish it from the 626-aa form called ALF4) is identical to ALF4 for the first 245 aa; however, aa 246 is a valine in ALF4ΔC compared to an asparagine in ALF4.

ALF4 is expressed in tissues affected by the alf4-1 mutant and its encoded protein is localized to the nucleus

To study the expression and subcellular distribution of ALF4, we used a β-glucuronidase (GUS) reporter in which GUS was fused in frame to aa 624 of ALF4. This fusion called ALF4(1-624)::GUS contains 2593 bp of genomic DNA 5′ to the predicted ALF4 translational start. Expression of ALF4(1-624)::GUS was observed throughout the root tip, stele, and lateral root primordia (Figure 3a). Expression was also observed in the shoot, whose development is also affected by the alf4-1 mutant (Figure 3a). Auxin (1 µm) treatment lasting for 1, 3, 6, 24, or 48 h had no effect on ALF4(1-624)::GUS expression with regard to tissue distribution or level of GUS activity (data not shown). In addition, examination of publicly available Arabidopsis microarray experiments is consistent with ALF4 not be regulated by auxin (available through, using microarray elements 144K1T7 and 98C10T7).

Figure 3.

Tissue-specific expression and subcellular localization of ALF4 reporters.

(a) Tissue localization of ALF4(1-624)::GUS. Representative plants showing GUS staining in the primary root, lateral root primordium, young seedlings, and leaves.

(b) Nuclear localization of ALF4(1-624)::GUS. The same root was stained for GUS activity and stained with DAPI. GUS staining was photographed under bright field, and DAPI was photographed by fluorescence microscopy. Arrows indicate corresponding nuclei.

(c) ALF4(1-246*)::GUS is excluded from the nucleus. ALF4(1-624)::GUS plants and ALF4(1-246*)::GUS plants were stained for GUS activity, and root epidermis cells in the elongation zone of the primary root are shown.

(d) Nuclear localization of ALF4(1–624)::GFP in the primary root tip. Root tip of an ALF4(1-624)::GFP plant visualized for GFP fluorescence and DIC. The merged image shows the GFP fluorescence in the nucleus but excluded from the nucleolus.

(e) Root cells of ALF4(1-624)::GFP plants stained with DAPI were photographed using GFP fluorescence, DAPI fluorescence, and DIC microscopy.

The ALF4(1-624)::GUS fusion protein contains almost the entire predicted ALF4-coding sequence and should therefore reflect the localization pattern of ALF4. As seen in Figure 3(b), ALF4(1-624)::GUS plants displayed a striking localization of GUS activity to nuclei (indicated by 4′,6-diamidino-2-phenylindole (DAPI) fluorescence) in root epidermal cells. Nuclear localization was less discernible in smaller cells close to the root tip where the nucleus makes up a large percentage of the cell and the X-Gluc precipitate interferes with the resolution of subcellular components.

A second reporter construct, ALF4(1-246*)::GUS, was designed to form a GUS fusion with ALF4 only when alternate splicing occurred leading to ALF4ΔC. This fusion at +1519 bp from the start of translation in the ALF4 genomic sequence will yield the ALF4(1-246*)::GUS fusion protein only when alternately spliced; otherwise, GUS would be out of frame with ALF4. Transgenic Arabidopsis carrying ALF4(1-246*)::GUS shows GUS activity with the same tissue distribution as that in the case of ALF4(1-624)::GUS, although GUS activity was somewhat reduced in leaves (data not shown). As compared to ALF4(1-624)::GUS, ALF4(1-246*)::GUS was mostly cytoplasmic in the root epidermis cells. In fact, partial nuclear exclusion of GUS was evident for ALF4(1-246*)::GUS, suggesting that the signal for nuclear localization is contained within the C-terminal 380 aa of ALF4 (Figure 3c). These results indicate that ALF4ΔC is synthesized in planta. Its different intracellular location suggests that this shorter form may have a function distinct from that of the 626-aa form.

To confirm the nuclear localization of ALF4, a fusion similar to ALF4(1-624)::GUS was constructed except that green fluorescent protein (GFP) was substituted for GUS. Transgenic plants carrying ALF4(1-624)::GFP show clear nuclear staining in all cells in the root in which it is expressed, including the small cells of the root tip where nuclear localization was unclear using the GUS reporter (Figure 3d). In root epidermis cells, as with ALF4(1-624)::GUS, co-localization of the reporter with DAPI was seen (Figure 3e). In some cases, especially in cells closer to the root tip, ALF4(1-624)::GFP appeared to have limited access to the nucleolus indicated by the non-fluorescent center of the nucleus (Figure 3d).

ALF4(1-624)::GFP expression like that of ALF4(1-624)::GUS was not altered by auxin treatment. In addition, no change in ALF4(1-624)::GFP intracellular localization was seen after auxin treatment (data not shown).

ALF4 is required for 35S::NAC1 induction of lateral roots

The alf4-1 defect in lateral root development is not rescued appreciably by exogenous IAA or excess IAA produced by the alf1-1 allele of ROOTY (Celenza et al., 1995) or the yucca (Zhao et al., 2001) IAA-overproducing mutant (data not shown). By cloning the ALF4 gene, we have now determined that ALF4 is expressed in most plant tissues, and its expression is not regulated by IAA. Taken together, these results suggest that ALF4's role in lateral root formation is not in auxin metabolism and may be independent of auxin signaling. To test the role of ALF4 in auxin-signaling pathways, we analyzed 35S::NAC1 alf4-1 double mutants. NAC1 encodes a transcription factor thought to function in auxin signaling downstream of the primary auxin response (Xie et al., 2000). Transgenic plants overexpressing NAC1 (35S::NAC1) exhibit excess lateral root formation, and 35S::NAC1 suppresses the lateral root defect caused by the transport inhibitor response 1 (tir1) mutation (Xie et al., 2000). 35S::NAC1 plants constructed in the Landsberg erecta (Ler) background were crossed to alf4-1/ALF4 plants (Columbia (Col-0) background), and alf4-1/ALF4 F2 plants were identified that either were homozygous for 35S::NAC1 or did not contain the transgene. In progeny derived from either F2 class, the alf4-1 mutation segregated in the expected Mendelian ratio. Plants from the F2 progeny show a strong alf4 mutant phenotype whether 35S::NAC1 is present or not (Figure 4). This result indicates that ALF4 is required for 35S::NAC1 to exert its effect on lateral root formation and suggests that if ALF4 is part of an auxin-signaling pathway, it likely functions downstream of NAC1. Alternatively, ALF4 could function in a developmental pathway independent of auxin signaling.

Figure 4.

alf4-1 is epistatic to 35S::NAC1.

35S::NAC1 plants constructed in the Ler background were crossed to alf4-1/ALF4 plants in the Col-0 background. alf4-1/ALF4. F2 plants were identified that were homozygous for 35S::NAC1 or that did not contain the transgene. From these two classes of plants, wild-type (WT), 35S::NAC1, 35S::NAC1 alf4-1, and alf4-1 plants were identified. Plants were photographed after 2-week growth.

Xylem-adjacent pericycle is present in the alf4-1 mutant

If ALF4 acts independently from the auxin signal/reception pathway, it could affect the establishment or maintenance of the xylem-adjacent pericycle. The GFP-expressing enhancer trap line J0121 marks the xylem-adjacent pericycle cells in wild-type plants (Casimiro et al., 2001). This reporter is expressed in the xylem-adjacent pericycle as it forms in the root apical meristem, and therefore serves as an early indicator for the specification of the lateral-root competent pericycle cells. We crossed J0121 to the alf4-1 mutant and found that in alf4-1, the reporter is expressed in the xylem-adjacent pericycle cells as it is in the wild type (Figure 5). As J0121 is expressed normally in the alf4-1 mutant, the mutant appears to form xylem-adjacent pericycle cell files and acts downstream of the initial formation of these cells.

Figure 5.

Expression of root tissue-specific marker J0121 in the alf4-1 mutant.

Wild-type and alf4-1 mutant progeny from strains homozygous for the J0121 reporter and heterozygous for the alf4-1 mutation were photographed using GFP fluorescence and autofluorescence (red) after which the images were merged.

alf4-1 pericycle cells are defective in completing mitosis

Although the xylem-adjacent pericycle is established in the alf4-1 mutant, it might not be maintained in a state competent for lateral root formation. A hallmark of the xylem-adjacent pericycle cells is that these cells remain mitotically active after leaving the meristem (Dubrovsky et al., 2000). To determine the mitotic behavior of the pericycle cells, we examined the expression of GUS reporters for the cyclin-dependent kinases CDKA;1 and CDKB;1 and the mitotic cyclin CYCB1;1 in the alf4-1 mutant.

Expression of CDKA;1 correlates with a general competence for cell division, and neither its expression is affected by auxin nor is it cell cycle regulated (Hemerly et al., 1993; Martinez et al., 1992). Expression of CDKA;1::GUS appears normal in alf4-1 plants compared to the wild type, being highest in the primary root tip and extending along the stele of the primary root (Figure 6a). This expression pattern supports a model in which alf4-1 pericycle cells maintain a general competence for cell division. However, as CDKA;1 expression itself is not cell cycle regulated, its expression does not offer a measure of the extent to which alf4-1 pericycle cells progress through the cell cycle.

Figure 6.

Expression of CDKA;1::GUS, CYCB1;1::GUS, and CDKB;1::GUS in the alf4-1 mutant.

Wild-type and alf4-1 seedlings carrying CDKA;1::GUS(a), CYCB1;1::GUS(b), or CDKB;1::GUS (c) reporters were grown in liquid PNS for 10 days and then assayed for GUS activity for 1 h (CDKA;1::GUS), 6 h (CYCB1;1::GUS), or 48 h (CDKB;1::GUS) at 37°C.

Expression of the mitotic cyclin CYCB1;1 is indicative of dividing pericycle cells and correlates with the initiation and subsequent cell divisions that occur during the development of lateral root primordia (Ferreira et al., 1994). If alf4-1 plants are defective in the initiation of lateral roots, expression of CYCB1;1::GUS should be absent in the xylem-adjacent pericycle in alf4-1 plants. The CYCB1;1::GUS reporter we constructed contains the first 150 aa of CYCB1;1 fused to GUS creating a destabilized GUS enzyme. Thus, CYCB1;1::GUS only stains actively dividing cells or cells arrested in mitosis. (This reporter is similar to one described by Colón-Carmona et al., 1999). As expected, wild-type plants show CYCB1;1::GUS expression in the primary root tip and at sites of lateral root initiation (Figure 6b). However, in alf4-1 plants, expression is found only in the root tip (Figure 6b), consistent with alf4-1 xylem-adjacent pericycle cells being unable to undergo the cell divisions needed to initiate lateral roots. While CYCB1;1::GUS expression is completely absent in the alf4-1 pericycle, expression is present, although reduced, in the root tip of the alf4-1 mutant compared to the wild type (Figure 6b).

The absence of CYCB1;1::GUS expression in the pericycle of alf4-1 plants suggests that alf4-1 pericycle cells may be defective in cell cycle progression at a step prior to mitosis. CDKB;1 is expressed earlier in the cell cycle than CYCB1;1 (Segers et al., 1996). Therefore, we examined the expression of CDKB;1::GUS in alf4-1 plants. In wild-type plants, CDKB;1::GUS is expressed in the primary root meristem and weakly in the pericycle and very young lateral root primordia (Figure 6b). alf4-1 plants show an overall increase in CDKB;1::GUS expression within the root meristem (Figure 6b) and this increase extends beyond the meristem into the pericycle. Increased CDKB;1::GUS expression in alf4-1 roots suggests that CDKB;1 accumulates in these tissues, possibly because cells are blocked or delayed in the cell cycle at a step earlier than the point at which CYCB1;1 is expressed. Thus, alf4-1 root tissues, compared to wild-type root tissues, overexpress CDKB;1::GUS and fail to express CYCB1;1::GUS.

Auxin rescue of the alf4-1 mutant is limited to newly formed pericycle cells

Prolonged IAA treatment of the alf4-1 mutant results in the occasional formation of a lateral root in the young part of the root, suggesting that IAA, a powerful mitogen, can promote limited cell division in the alf4-1 mutant (Celenza et al., 1995). Therefore, we examined the effect of IAA on expression of CYCB1;1::GUS and CDKB;1::GUS in the alf4-1 mutant. After a 72-h treatment with 1 µm IAA, one to three lateral roots did initiate in the first few millimeters of alf4-1 roots and these lateral root primordia show both CYCB1;1::GUS and CDKB;1::GUS expression (Figure 7a). The area on alf4-1 plants where lateral roots form corresponds to what would have been, prior to IAA treatment, the elongation zone directly behind the root apical meristem. These primordia that form on alf4-1 are less developed than those that form on similarly treated wild-type plants, suggesting that development is delayed in alf4-1 plants compared to wild-type plants.

Figure 7.

Auxin treatment can partially rescue lateral root formation on the alf4 mutant.

(a) Auxin treatment for 72 h induces lateral roots on the younger portion of alf4-1 roots. Wild-type and alf4-1 seedlings carrying the CYCB1;1::GUS or CDKB;1::GUS reporter were grown in liquid PNS for 7 days after germination and then treated with 1 µm IAA for 72 h. GUS activity was assayed for 6 h (CYCB1;1::GUS) or 48 h (CDKB;1::GUS) at 37°C.

(b) Auxin treatment for 24 h induces CYCB1;1::GUS expression only in the younger pericycle cells of alf4-1 roots. Wild-type and alf4-1 seedlings carrying the CYCB1;1::GUS reporter were grown in liquid PNS for 7 days after germination and then treated with 1 µm IAA for 24 h. GUS activity was assayed for 6 h. Composite images were assembled using adobe photoshop version 5.0. Bar indicates 1 mm.

(c) Lateral roots induced on alf4-1 roots are less developed than similarly treated wild-type roots. Wild-type and alf4-1 seedlings carrying the CDKB;1::GUS reporter were grown in liquid PNS for 7 days after germination and then treated with 1 µm IAA for 24 h. GUS activity was assayed for 48 h.

Because of these results, we next examined CYCB1;1::GUS and CDKB;1::GUS expression in alf4-1 pericycle cells treated with IAA for 24 h to determine if IAA induction of cell division was more extensive earlier in development. Indeed, we did detect more extensive pericycle expression of CYCB1;1::GUS in alf4-1 plants exposed to 1 µm IAA for 24 h compared to 72 h of treatment (Figure 7b). However, these newly initiated lateral roots were limited to the first 8 mm of alf4-1 plants whereas similarly treated wild-type plants displayed CYCB1;1::GUS expression in lateral root primordia along the entire root (Figure 7b). In addition, while lateral root primordia formed in wild-type plants were typically at developmental stages III–V, alf4-1 primordia were only at stages I–III (Figure 7c). For CDKB;1::GUS, we again observed increased reporter expression in the alf4-1 mutant even without IAA. Interestingly, we found that the lateral root primordia that formed on alf4-1 roots after IAA treatment had much stronger CDKB;1::GUS expression than the expression found in wild-type lateral root primordia formed in similarly treated plants (Figure 7c).

In the course of doing these experiments, we noted that the number of primordia (stage I or older) observed on wild-type roots after 24 h of IAA treatment (19 ± 1.9) was equivalent to the number of emerged lateral roots observed after a 72-h IAA treatment (20 ± 1.3; Figure 8). However, the number of primordia found on alf4-1 roots after a 24-h IAA treatment (6.4 ± 1.4) was several fold greater than the number of emerged lateral roots found after a 72-h IAA treatment (2.3 ± 0.31; Figure 8). This result is consistent with IAA being able to rescue only partially the alf4-1 defect in terms of both location and total number of lateral roots initiated as well as the number of primordia that eventually mature.

Figure 8.

Auxin induces fewer lateral roots on alf4-1 plants.

Quantification of stage I or older lateral root primordia (24 h) or emerged lateral roots (72 h) was performed on wild-type or alf4-1 plants carrying the CDKB;1::GUS reporter treated with 1 µm IAA for 24 or 72 h after growth for 7 days post-germination. For each condition and genotype tested, numbers of lateral roots from eight plants were averaged and SEs were calculated.


The cloning of the ALF4 gene has provided a clearer understanding of its role in lateral root formation. Our finding that ALF4 is a plant-specific gene suggests that it has a role unique to plant development. The phenotype of alf4 mutants, premature loss of mitotic potential in the xylem-adjacent pericycle, defines genetically the role of the pericycle as the focal point of root system developmental plasticity, a function consistent with the plant-specific origin of this gene.

Using the cloned gene, we have found that ALF4's expression and intracellular location is not regulated by auxin. This finding, combined with the discovery that ALF4 is required for 35S::NAC1 to induce excess lateral roots, suggests that if ALF4 were to function in auxin signaling, it would have to act downstream of NAC1. In addition, our finding that the xylem-adjacent pericycle-specific reporter J0121 is expressed in the alf4-1 mutant, indicates that these cells are formed in the mutant. Because of these results, we favor an alternative explanation: ALF4 is required for maintaining the meristem-like properties of the xylem-adjacent pericycle that allow the these cells to divide and re-differentiate into a lateral root primordium. According to this hypothesis, signals for the induction of lateral roots, such as auxin, can be perceived in the alf4-1 mutant, but lateral root formation cannot proceed because the pericycle cells are impaired in their ability to divide in response to these signals. Consistent with this model, we found that the alf4-1 mutant failed to express CYCB1;1::GUS in the pericycle, indicating that ALF4 is required to maintain the xylem-adjacent pericycle cells in a mitotically active state needed for lateral root formation. That auxin can only partially remediate lateral root formation in the alf4-1 mutant is also consistent with this hypothesis; auxin is able to weakly sustain xylem-adjacent pericycle cell division in the absence of ALF4, as long as the cell is actually dividing. However, once cell division has ceased, auxin cannot re-activate the cell cycle unless ALF4 is present, suggesting that ALF4's primary role may be to prevent terminal differentiation of the xylem-adjacent pericycle. In support of this model, we have found that alf4-1 roots are greatly reduced in their ability to form callus (data not shown). In addition, in alf4-1, pericycle cell division is greatly reduced as compared to wild-type roots when seedlings are grown in conditions designed to synchronize pericycle cell division in response to auxin (Himanen et al., 2002; data not shown).

Recent studies on root morphogenesis are consistent with a model in which the xylem-adjacent pericycle needs to be maintained in a mitotically competent state. These analyses suggest that xylem-adjacent pericycle cells (the pericycle files that can form lateral roots) continue to divide for some time after leaving the meristem, creating what can be considered an extended meristem, and that these cells do not terminally differentiate (Beeckman et al., 2001; Dubrovsky et al., 2000). The simplest explanation of the alf4-1 phenotype is that xylem-adjacent pericycle cells do terminally differentiate and stop dividing earlier than the wild type; thus, lateral roots do not form.

Treatment with IAA does not induce lateral root formation in the older regions of alf4-1 roots whereas these same regions of wild-type roots respond robustly to treatment with the hormone. This failure of alf4-1 plants to produce lateral roots correlates with the absence of induction of CYCB1;1. Prolonged IAA treatment can induce lateral roots to form in the youngest portions of the alf4-1 mutant, and in these cells, CYCB1;1 is expressed. The difference in mitotic potential between young and old pericycle cells may depend simply on their distance from the meristem or may depend on the relationship of these pericycle files to the xylem or other tissues. As only xylem-adjacent pericycle cells form lateral roots, it is tempting to postulate a signaling mechanism between the xylem and pericycle that is needed to confer or maintain the meristem-like properties of the xylem-adjacent pericycle. If such communication exists, ALF4 could function in this process. However, we note that when the occasional lateral root does form on the alf4-1 mutant, it still forms from the xylem-adjacent pericycle, indicating that the mechanisms that specify this tissue are still in place in the absence of ALF4.

Occasionally, IAA-induced alf4-1 primordia will emerge as mature lateral roots that become independent of exogenous IAA once a meristem forms (see Figure 7). Thus, we can refine our model such that ALF4 functions to maintain mitotic competence in pericycle cells that have left the meristem, but is not required to maintain the meristem itself. Nevertheless, we have observed less cell division in the primary root tip of the alf4-1 mutant, suggesting that ALF4 may function to maintain mitotic competence in other root tissues besides the pericycle.

How does ALF4 keep xylem-adjacent pericycle cells mitotically active? ALF4 could regulate the cell cycle directly, or alternatively, it could delay or prevent terminal differentiation in the xylem-adjacent pericycle, thus impacting the cell cycle indirectly. Delay of terminal differentiation would maintain tissues in a developmentally responsive state capable of mitosis. While we have focused on ALF4's role in lateral root development, shoot phenotypes are also seen in alf4 mutants (Celenza et al., 1995). Consistent with the mutant's pleiotropy, ALF4 reporter expression is seen in shoot tissues in addition to root tissues. Some of these shoot phenotypes, such as reduced hypocotyl elongation and leaf expansion, are likely not because of defects in cell division. Considering these phenotypes in combination with alf4-1's root phenotype, we suggest that the more general role of ALF4 is to preserve developmental plasticity in non-meristematic tissues via an as of yet undefined mechanism; in xylem-adjacent pericycle cells, plasticity is manifested most clearly by the ability to divide. As we learn more about ALF4's role in root development, we expect to gain insight into its shoot functions as well.

The availability of the cloned gene resolves some issues and raises others. Our interpretation of our results is based on the assumption that alf4-1 is a null allele. Indeed, the structure of the gene coupled with the analysis of alf4-1 mRNAs supports this hypothesis because alf4-1 transcripts would produce a 69-aa protein containing amino acids identical to only the first 55 aa of ALF4. In addition, the T-DNA insertion allele of ALF4 (alf4-063) has a phenotype identical to that of alf4-1.

The nuclear localization of protein fusions of ALF4 to GUS or GFP is suggestive of a regulatory role. Although a nuclear import sequence was not predicted computationally, the C-terminal 380 aa appear to be required for nuclear localization because ALF4(1-246*)::GUS is excluded from the nucleus. ALF4's nuclear location is consistent with a role in cell cycle regulation as well as more general roles such as transcriptional regulation; however, no similarities to proteins involved in these activities have been found.

The function of the ALF4ΔC is unclear, although its different subcellular location suggests a role distinct from the nuclear form. Fortuitously, the T-DNA insertion site of the alf4-063 allele is at the end of the fifth exon and would encode a truncated version of ALF4 that contained only the first 244 aa of ALF4 before extending into the T-DNA right border. Because of the similarity in primary structure of ALF4ΔC to the protein encoded by alf4-063 allele, it is likely that nuclear-localized ALF4 is needed for full ALF4 function. Nonetheless, for a complete understanding of the ALF4 gene, it will be important in the future to determine the functional significance of the different forms of the ALF4 protein. One possibility is that ALF4ΔC's role is to regulate the amount of full-length nuclear ALF4 perhaps by sequestering it in the cytoplasm through heterodimerization. Alternately spliced mRNAs leading to proteins with different subcellular locations are not unprecedented. For example, alternate splicing of the message encoding the antiapoptotic protein survivin results in one form that shuttles between the nucleus and cytoplasm and another form that is retained in the nucleus (Rodríguez et al., 2002).

Our cloning of the ALF4 gene and characterization of the alf4 mutant provide genetic evidence that the xylem-adjacent pericycle, while formally a differentiated tissue, maintains a subset of properties similar to those of meristem cells, including the ability to continue cell division and re-differentiate into a lateral root. ALF4 is required for the maintenance of these properties that in turn enable the xylem-adjacent pericycle to create the developmental plasticity needed to alter root system architecture. We speculate that the appearance of the ALF4 gene in plants coincides with the ability of differentiated plant cells to modify their developmental potential in response to intrinsic or extrinsic signals.

Experimental procedures

Plant strains and growth conditions

The Col-0 ecotype was used for construction of all transgenic plants and double mutants unless otherwise stated. 35S::NAC1 seeds were a gift from N.-H. Chua.

Seeds were germinated under sterile conditions in plant nutrient sucrose (PNS) medium (Haughn and Somerville, 1986). For solid PNS medium, agar (0.6%) was added. Transformation of plants was performed by floral dip (Clough and Bent, 1998), and for selection of transgenic plants, PNS medium without sucrose containing either kanamycin (15 µg ml−1) or hygromycin (15 µg ml−1) was used. Plates were sealed with gas-permeable Micropore tape (3M Health Care, St Paul, MN, USA), and then incubated at 21°C under constant light using yellow low-pass filters (Stasinopoulos and Hangarter, 1990) at an intensity of 20–30 µE m−2 sec−1 unless otherwise noted.

Recombinant DNA methods

Unless noted otherwise, standard microbiological and recombinant DNA techniques were employed for growth of bacterial strains, restriction digestion analyses, ligations, and subcloning (Ausubel et al., 1994–1998). DNA sequencing was performed at the Boston University Sequencing Facility using an ABI Prism® 377 DNA Sequencer (Applied Biosystems, Foster City, CA, USA), and sequence analysis was performed using ABI Prism editview Version 1.0.1 software (Applied Biosystems, Foster City, CA, USA) or lasergene navigator software (DNAStar Ltd., London, UK).

PCR were performed in 50-µl volumes for 40 cycles (94°C for 1 min, 58°C for 1 min, and 72°C for 3 min) followed by incubation at 72°C for 10 min. Oligonucleotides were obtained from Genemed Synthesis (South San Francisco, CA, USA) or Invitrogen (Carlsbad, CA, USA).

Positional cloning of ALF4

Details of the positional cloning of ALF4 are provided in Supplementary Material.

Bioinformatic methods

blastn, blastx, and tblastn analysis ( of EST, cDNA, and genomic sequence was conducted using the default E-value cutoff of 1.0 and default algorithm setting BLOSUM62 as previously described by Altschul et al. (1997). Sequence alignments and protein statistics were determined using lasergene navigator version 1.60 software and San Diego Super Computer Center (SDSC) Biology Workbench ( Kyte-Doolittle analysis was conducted using grease, a hydropathy program available through the SDSC Biology Workbench (Kyte and Doolittle, 1982). psort ii ( (Nakai, 2000; Nakai and Kanehisa, 1992) was used to analyze ALF4 for the presence of intracellular targeting sequences. For identification of sequences corresponding to protein modification sites, scanprosite (http:/, including and excluding commonly occurring sequences, was used (Falquet et al., 2002). pfam-a ( (Bateman et al., 1999)) and block searcher ( (Henikoff and Henikoff, 1994)) were used to search for protein domains in ALF4.

Identification of alf4-063

A putative ALF4 T-DNA insertion mutant, SALK_063183, was identified from the Salk Institute Genomic Analysis Laboratory (SIGnAL) T-DNA insertion collection ( and obtained from the Arabidopsis Biological Resource Center (ABRC). Seeds were sterilized and grown on PNS medium; approximately 20% of the seeds showed an alf4-1 phenotype. The T-DNA left-border insertion junction with ALF4 was determined by PCR amplification of the junction with primers LBa1 (5′-tggttcacgtagtgggccatcg-3′; and ALF4R063 (5′-tgttatgtcaagcaagaactcgat-3′). The amplification product was sequenced, and the insertion site was determined to be between bases +1900 and +1901 of the ALF4 genomic sequence. This allele of ALF4 was named alf4-063.


For RT-PCR of ALF4 mRNAs, total RNA was isolated from wild-type or alf4-1 plants using the RNAqueous Kit (Ambion, Austin, TX, USA). RT was conducted with 1 µg of total mRNA using the Promega Reverse Transcription System (Promega, Madison, WI, USA). A 5′ primer beginning at −40 bp from the ALF4 initiation codon (5′-ccgggggatccgactatccggggttaaactctattc-3′) and a 3′ primer beginning at +1881 bp of the predicted ALF4 cDNA (5′-ccccggatccctaatgacttttcaacttttcttccac-3′) were used to PCR amplify ALF4 transcripts from cDNA. PCR was performed for 40 cycles (94°C for 1 min, 58°C for 1 min, and 72°C for 3 min). Products derived from these reactions were digested with BamHI (sites underlined in the primers), subcloned with BamHI into pBluescript II KS+ (Stratagene, La Jolla, CA, USA), and sequenced.

Reporter constructs

For construction of ALF4(1-624)::GUS, PCR was used to introduce a BamHI site into the ALF4 genomic sequence at position +4146 relative to the ALF4 initiation codon. Using this introduced BamHI site and a vector BamHI site, an ALF4 genomic fragment extending from −2593 to +4146 was cloned in frame to GUS in pCAMBIA1391Xa (CAMBIA, Canberra, Australia).

For the ALF41-246*::GUS construct, a BamHI site was introduced at position +1520 relative to the ALF4 start codon. Using this introduced BamHI site and a vector BamHI site, an ALF4 genomic fragment extending from −2593 to +1520 was cloned into pCAMBIA1391Xc (CAMBIA, Canberra, Australia). This fusion was designed to express GUS only when the alternative splice event occurs that adds four bases to exon five.

For ALF4(1-624)::GFP, the same ALF4 gene fragment used for ALF4(1-624)::GUS was subcloned into a GFP fusion vector constructed as follows. Plasmid psmGFP (Davis and Vierstra, 1998) was obtained from the ABRC and smGFP was amplified with primers 5′-tctagaggatcccaaaggagatataacaatgag-3′ and 5′-aggaaacagctatgaccatgattac-3′. The amplification product was digested with EcoRI and BamHI and subcloned into pPZP212 (Hajdukiewicz et al., 1994), also digested with EcoRI and BamHI to create the vector used to construct ALF4(1-624)::GFP.

CYCB1;1::GUS, which has been described previously by Donnelly et al. (1999), was constructed by subcloning into the BamHI–SmaI of pBI101.3 (Clontech, Palo Alto, CA, USA) a genomic fragment of CYCB1;1 beginning at the BamHI site, −3050 bp from the initiation codon and extending to the EcoRV site at +908. EcoRV cleaves after codon 150 of CYCB1;1 and creates an in-frame fusion with GUS.

CDKA;1::GUS was constructed by subcloning cloning into the XbaI site of pBI101.2, a genomic fragment of CDKA;1 beginning −3236 bp from the initiation codon and extending to an artificial XbaI site added immediately after the initiation codon.

CDKB;1::GUS was constructed by subcloning into the XbaI site of pBI101.2, a genomic fragment of CDKB;1 beginning approximately −4000 bp from the initiation codon and extending to an artificial XbaI site added immediately after the initiation codon.

J0121 has been described previously ( (Casimiro et al., 2001)) and was obtained from the ABRC.

Reporter analyses

Wild-type and alf4-1 mutant seeds carrying various GUS and GFP reporters were grown hydroponically in 3 ml of PNS at 21°C under continuous light in six-well culture dishes for 7–10 days post-germination. For IAA induction experiments of CYCB1;1::GUS and CDKB;1::GUS plants, seedlings were grown hydroponically for 7 days post-germination, after which IAA was added to the medium (final concentration 1 µm) and plants were grown for an additional 24 or 72 h.

Quantification of lateral root primordia (stage I or older) was performed on eight wild-type or alf4-1 plants carrying the CDKB;1::GUS reporter grown in 1 µm IAA for 24 or 72 h. For each condition and genotype tested, numbers of lateral root primordia were averaged and SEs were calculated.

To assay GUS activity, PNS medium was removed and replaced with 3 ml of GUS buffer (0.1 m NaPO4, 0.5 mm K3[Fe(CN6)], 0.5 mm K4[Fe(CN6)], 10 mm EDTA, 0.01% Triton X-100, and 0.25 mg ml−1 5-bromo-4-chloro-3-indolyl β-d-glucuronide (Rose Scientific, Ltd, Edmonton, Canada) dissolved in 100 µl of dimethyl formamide) and incubated at 37°C (Jefferson et al., 1987). Assay times were 12–36 h for plants carrying ALF4(1-624)::GUS and ALF4(1-246*)::GUS constructs, 6 h for CYCB1;1:: GUS, 1 h for CDKA;1::GUS, and 48 h for CDKB;1::GUS. GUS buffer was then removed, and plants were rinsed once in sterile H2O, fixed in 70% ethanol, and stored in 50% glycerol and 0.01% Triton X-100 prior to microscopy.

To confirm the ALF4 genotype of GUS-stained plants, shoot tissue was analyzed by PCR using primers 5′-gctttgagatcgatgaatactta-3′ and 5′-aactctcaaagtcttgaaatcct-3′, which amplify a 123-bp fragment for wild-type ALF4 allele and a 111-bp fragment for the alf4-1 allele.

Where indicated DAPI staining was carried out by incubating plants for 10 min in sterile H2O (ALF4(1-624)::GFP) or in 50% glycerol and 0.01% Triton X-100 (ALF4(1-624)::GUS) containing 5 µg ml−1 DAPI.


For observation of GUS expression, stained plants were examined by light microscopy using a Leica MZ6 dissection scope (Leica Microsystems, Heidelberg, Germany) or by differential interference contrast (DIC) using a Nikon Eclipse E600 microscope (Nikon Inc., USA) with a mounted Pixera VCS digital camera and pixera studio version 1.2 software (Pixera Corporation, Los Gatos, CA, USA). A Zeiss Axioplan fluorescent scope (Carl Zeiss Inc., Thornwood, NY, USA), courtesy of Dr C. Li (Boston University), was used for fluorescent microscopy to visualize DAPI-stained roots.

GFP fluorescence was visualized on an Olympus BX50 laser scanning fluorescence microscope using 488 nm for GFP and 568 nm for red autofluorescence, and transmitted for DIC. Images were obtained using the Olympus Fluoview Digital Image Capture System with olympus fluoview software 2.1.37. Images were processed with adobe photoshop 5.0.


We thank the ABRC for strains, N.-H. Chua for the 35S::NAC1 strain, S.D. Michaels for the XY CAPS primer sequences, and M. Stammers for providing YAC information. We also thank A.K. Hull and J. Hickey for assistance in library screening. This work was supported by a National Science Foundation grants IBN-9514096 (J.L.C.) and MCB-9974451 (G.R.F.) as well as the Department of Biology and Undergraduate Research Opportunities Program at Boston University.

Supplementary Material

The following material is available from

Details about genetic markers and clones used in the positional cloning of ALF4 as well as a map of the locus are available as a separate PDF file.